Fundamental Thinking

Animated GIF for Presentations

2012-01-18T09:05:00.000-08:00

If you're using powerpoint for presentations you know how tricky animations can be if you're presenting on different computers. PPT 2010 seems to be better at handling the problem of embedding files but I don't know. PDF can now play video but it'll probably be awhile before the world as a whole can use it. That means LaTeX and beamer might not be useful yet (I LOVE beamer presentations though!).

I've mentioned before that I work in an office with both Mac and PC and a variety of PPT versions so I need compatibility across. I also want decent resolution without massive files attached to my presentations so I'm converting my .avi to animated .gif. These files seem to be about 1/10 to 1/5 the file size and SHOULD embed easily into a powerpoint document.

I'm not going to write out the steps to making nice animated .gif files because this guy (xFL1PPYx) did such a WONDERFUL job of showing how to do it. He uses GIMP and GAP (GIMP Animation Package). Both are free. The links are given in his video description.

Check it out!

http://www.youtube.com/watch?v=u5_3MGP2Lj4&feature=related

Generating Publication Quality Figures in Matlab

2012-01-13T04:46:00.000-08:00

Updated* See Step 7 - Post Processing

Matlab is wonderful for visualizing solutions quickly for the purposes of debugging or exploring numerical solutions. It is pretty bad at figure formatting though. In fact, I don't feel comfortable putting Matlab figures directly into documents because they look so unprofessional. Matlab does have the workspace toolbox thing, but that's not the best. Mostly because the export is exactly the same size as the figure on the screen and setting that size is really inconvenient so axis of adjacent figures don't usually line up well.

This is how I do it:

First we are assuming either the correct sizing for a figure that will fit on half a page so that 2 can be side by side on a 6.25 (or 6.5) inch text width OR that I have one elongated figure that fills the whole width - like a contour plot.

1) Specify the dimensions

% For Normal Figures

height=1.0/1.618; % width/golden ratio

width=1;

% For Wide Figures

% height=1.0/1.618; % width/golden ratio

% width=2;

scale=300; % 3.13 inches

I choose the golden ratio because it's considered aesthetically pleasing to the eye. It works well for large figures, but it might not be perfect for small figures. Just remember if you want to scale from a known dimension in inches use a converter to convert to pixels. 300 pixels is 3.13 inches

You can also position the figure on the screen

xpos=50;

ypos=500;

2) Define the function to be plotted. For the example, I generate one on the spot but this could be an import function

x=0:.01:2*pi;

f1=cos(x);

f2=sin(x);

3) Generate the figure. This doesn't just mean plot the data. The figure is comprised of the figure dimensions, the plot area, the bounding box, etc.. We also want to set the fonts and position

figure; % Create Figure

axes('FontName','Times New Roman') % Set axis font style

box('on'); % Define box around whole figure

set(gcf,'Position',[xpos ypos scale*width scale*height]) % Set figure format

4) Plot the data

hold on

plot1=plot(x,f1,'Color',[1 0 0]);

plot2=plot(x,f2,'Color',[0 0 1]);

by plotting each function as a different plot command and defining plot1 and plot2, we have unique control over the format of each data set.

5) Set Plot properties. This is different than setting figure properties and refers to the data set format

set(plot1,'LineWidth',1,'LineStyle','-');

set(plot2,'LineWidth',1,'LineStyle','--');

% Set Axis Limits

xlim([min(x), max(x)])

ylim([min(f1), max(f1)])

% Create xlabel

xlabel('\xi','FontSize',11,'FontName','Times New Roman','FontAngle','italic');

% Create ylabel

ylabel('\eta','FontSize',11,'FontName','Times New Roman','FontAngle','italic','rot',0);

% Create Legend

hleg1 = legend('$\cos(x)$','$\sin(x)$');

% Set Legend Properties

set(hleg1,'Interpreter','latex')

set(hleg1,'Location','SouthWest')

set(hleg1,'box','on')

There are more properties that can be set, but i just took the ones I use most.

6) Export figure

fig = gcf;

style = hgexport('factorystyle');

style.Bounds = 'loose';

hgexport(fig,'Example_Figure.eps',style,'applystyle', true);

drawnow;

print -depsc2 -tiff myfile.eps

There is something weird here: i don't think i have to use both the hgexport (like clicking File-Save As) and print but I don't get the right format without using both.

7) Post processing. This is really unfortunate. The problem is that Matlab doesn't embed fonts in eps files correctly (or at all). There are functions such as export_fig and exportfig that claim to do this, but I've not had any luck. Part of it is that I don't have time to mess with all the settings and syntax that comes with using other packages. Part of it is my frustration with the whole nonsense.

An option many people seem to use is Adobe Illustrator. The process goes: Open the eps and convert the text to outlines. I think this replaces the text with lines and fills so that the letters appear, but aren't actually text anymore. That's a good idea, but I spent some time poking around the trial and couldn't figure out how to set the damn page dimensions so the exported eps was back to its original size. It would always export it with a big white space around it. I think there's something about the clipping box but I don't have time to mess with this garbage. I'm a scientist, not a graphic designer.

Next option: ACD Canvas. Open the eps in canvas and convert to canvas object. Magically (expected), it imports the eps figure with the correct dimensions! Then select all and "convert to path." I'm not completely sure what goes on behind the scenes with this operation, but it appears to take whatever is selected and lump it all into some kind of vector graphic. Again, the text is no longer dependent on a font. The problem with this is that there is no more editing so make sure it's what you want before you convert it. This works fine and produces nice figures.

UPDATE - Canvas costs money beyond the trial version. Boo. Inkscape is the solution! It's a free, open source program that will do this stuff MUCH easier than Adobe and even easier as Canvas. Open the eps file. Make sure the fonts imported correctly. If not, fix them! Then save as eps. The dialog box will offer some really cool stuff for latex but we don't need that right now. Make sure the box "Convert texts to paths" is selected and the "export area is drawing" is selected (I don't know about the "export area is page" that sounds counter to the previous box but I left mine checked). I don't know what the rest does so leave it or not. It doesn't seem to matter. The important part is that the text is converted to paths!

Beyond that, Inkscape (and the others) let you edit figures. For instance, Matlab will automatically adjust the position of the axis and labels depending on the number of digits in the label. That means the axis won't necessarily line up correctly in the document. Inkscape can adjust the axis dimensions and the location of the labels and titles so that they all look correctly! Awesome!

This isn't perfect but it works pretty well. If you're without a better option, then this is definitely a viable possibility. Here is the whole matlab code:

%% Figure Generator with Format

% =========================================================================

clear;clc;close all

% =========================================================================

% Specify Dimensions and Position on Screen

% =========================================================================

% -------------------------------------------------------------------------

% Figure Dimensions

% -------------------------------------------------------------------------

% For Normal Figures

height=1.0/1.618; % width/golden ratio

width=1;

% For Wide Figures

% height=1.0/1.618; % width/golden ratio

% width=2;

scale=300; % 3.13 inches

% -------------------------------------------------------------------------

% Figure Position on Screen

% -------------------------------------------------------------------------

xpos=50;

ypos=500;

% =========================================================================

% Define Functions to be Plotted

% =========================================================================

x=0:.01:2*pi;

f1=cos(x);

f2=sin(x);

% =========================================================================

% Generate Figure

% =========================================================================

% -------------------------------------------------------------------------

% Figure Properties

% -------------------------------------------------------------------------

figure; % Create Figure

axes('FontName','Times New Roman') % Set axis font style

box('on'); % Define box around whole figure

set(gcf,'Position',[xpos ypos scale*width scale*height]) % Set figure format

% -------------------------------------------------------------------------

% Plot Data

% -------------------------------------------------------------------------

hold on

plot1=plot(x,f1,'Color',[1 0 0]);

plot2=plot(x,f2,'Color',[0 0 1]);

% -------------------------------------------------------------------------

% Plot Properties

% -------------------------------------------------------------------------

set(plot1,'LineWidth',1,'LineStyle','-');

set(plot2,'LineWidth',1,'LineStyle','--');

% Set Axis Limits

xlim([min(x), max(x)])

ylim([min(f1), max(f1)])

% Create xlabel

xlabel('\xi','FontSize',11,'FontName','Times New Roman','FontAngle','italic');

% xlabel('$\xi$','FontSize',11,'FontName','Times New Roman','interpreter','LaTex','rot',0);

% Create ylabel

ylabel('\eta','FontSize',11,'FontName','Times New Roman','FontAngle','italic','rot',0);

% ylabel('$\eta$','FontSize',11,'FontName','Times New Roman','interpreter','LaTex','rot',0);

% Create Legend

hleg1 = legend('$\cos(x)$','$\sin(x)$');

% Set Legend Properties

set(hleg1,'Interpreter','latex')

set(hleg1,'Location','SouthWest')

set(hleg1,'box','on')

% =========================================================================

% Export Figure

% =========================================================================

fig = gcf;

style = hgexport('factorystyle');

style.Bounds = 'loose';

hgexport(fig,'Example_Figure.eps',style,'applystyle', true);

drawnow;

print -depsc2 -tiff myfile.eps

Figure placement in Word

2011-11-11T06:29:00.000-08:00

Word handles figures pretty poorly for anything substantial. I've complained about it's inadequacies when writing scientific documents often.

Fixing figures to page locations is a bad idea when there are lots of them to put in because adding content above them will likely make any subsequent figures shift and overlap. This is kind of because they are anchored to a paragraph and a page location. Rather, anchoring them to a paragraph and allowing them to float with the text will work much better in the long run.

The best way, is to use frames although the necessity is going away with newer versions of word. Originally, frames ensured that the content within them appeared on the same layer as the content outside where text boxes floated on another layer. This makes cross-referencing impossible because the layers don't play nice. I'm not sure this is the case anymore though. In new versions of word you have to add the "Format Frame" button to the quick bar because it's not available by default.

Now, don't make frames and put figures into them. Do it the other way around. Put the figure into the document. Right click and insert caption. Then highlight the caption and the figure and click the format frame button. It will make a frame the correct height and the width of the page. Position the figure within the frame (i.e. center it) and you're done. The frame will be anchored to the next paragraph and will float nicely!

Removing field codes in Word Documents

2011-11-11T06:02:00.000-08:00

Field codes are used to create automatically generated content like cross-referencing, captioning, list of figures etc. Word is complete crap for technical writing and cross-platform collaboration but these things decent and should be used.

Why you would want to remove field codes? Beats the hell out of me. Just kidding. Submitting documents for publication might require fields removed. Cross-platform versions of word sometimes break the fields. Old versions of word do the same thing. Or if you're like me and you work with people who waste time by rewriting all the fields manually because they openly refuse to learn how to do it automatically. Seriously. I wouldn't care if it didn't waste my time but it always does.

Screw that nonsense. I found how to do both. I write up my reports the correct way - with fields - and strip the final version so that none exist and they can piss around doing whatever it is they do.

Thanks to this post, it's absolutely simple to do http://assorted-experience.blogspot.com/2008/04/remove-field-code-from-word-document.html

just like updating the whole document select all

ctrl-A

then strip the fields with

ctrl+shift+F9

Presto Manifesto!

Now that I can avoid intentionally doing things inefficiently, I feel much more motivated to finish this report. Thanks KG from the link!

Of course, there is a cross platform option that will always work...LaTex. Damn Word is crap for my work.

Unsorted reference numbering when using a list of figures and a citation is in the caption

2011-07-10T13:29:00.000-07:00

This is a dumb problem that has a little attention. I figured it out from

http://www.terminally-incoherent.com/blog/2006/07/26/latex-citations-in-figures/

If you use the cite command in the a caption AND have a list of figures or list of tables generated, then the unsorted numbering will be wrong. It will start numbering the references from the list of figures and not from the start of the actual document. Therefore, the first reference in the body of the document will not start at 1. To fix, use the optional command syntax for \caption:

\caption[This is the caption]{This is the caption \cite{REF}}

Just repeat the caption in the brackets and omit the \cite command. You'll probably have to rerun Latex twice and maybe bibtex once before the changes are updated but it works! The article I linked to says you might have to use the \protect{\cite{REF}} command, but I didn't find it necessary.

PhotoImpact: Stop Automatically Adjusting Image Size in Properties Menu

2011-03-18T09:14:00.000-07:00

Ok. So I tried PhotoImpact from ulead a long time ago. Not bad. They have a free trial and the full version starts at 50 dollars. That's a whole lot better than photoshop! I'm not good at image stuff but It was useful for making simple drawings. There was one problem I ran into and apparently others too.

Problem: When I make an object and size it, if I go back to the properties window the image has been resized by one or two pixels in some direction.

For most things, this isn't that big a deal and you'd never realize it. But if you need tight margins or exact positioning then this is awful! If I center a 50 by 50 image on a page (25 to the left of center and 25 to right) if it is resized to 50 by 51, then I can't center it if my original image has an even number of pixels.

Solution: Turn off anti-aliasing when creating the image. This option is in the tool properties menu. Click on the shape drawing tool. In the far right bottom of the top tool bar there is a button that looks like a window with check boxes and an arrow. This is the properties menu. Then in that menu uncheck anti-aliasing.

I get it, this feature makes lines smooth. But seriously, this is a terrible feature to have selected by default and hidden away with no explanation of what is going on with your figure.

All in all, I like this program. The learning curve is ok. The menus and settings are pretty damn awful but probably no different than photoshop. Best of all, it's cheap and does what it needs to do. If you don't need to get crazy with photoshop - or even if you do, you might want to see if this will work for you first.

Hyperlink Combo Boxes in Excel with VBA

2011-03-14T11:36:00.002-07:00

There is a lot of forums about how to do this and the best I've seen is using named ranges and data validation lists. I think there is a better way.

Problem: I want to make a drop down menu with hyperlinks to other worksheets

Solution: Write a simple VBA macro (I promise, it's really easy!)

Step 1: Make sure all your worksheets are cleanly titled. If you have several distinct groups of worksheets, then the first word of the title should be the same for the group. In this example, I have several worksheets grouped in "Tournament" so my titles are like "Tournament Setup", "Tournament Registration" etc.

Step 2: Create a worksheet to house all the menu options. This is not necessary if you can efficiently generate list items from the original worksheets. For example purposes, we will keep all our entries here. Also, if you want the code to work as shown below follow these steps exactly and name this sheet NavMenu.

Step 3: Designate a column for the menu items. If you want the code provided to work without modifying the cell locations, then the sheet should be built like

Cells B11:B17 are the menu items. For now, just type in the menu items in this range. The menu items should NOT include the group label. This will be added automatically in the code. Cell B8 is the group label. You can go ahead and add this too.

Step 4: Create a form control combobox (not activeX). Right click and click format control. In Input Range select the range containing the menu items. Here it is NavMenu!$B$11:$B$18. I include Cell B18 so I have a blank item to rest the menu to. You'll see what i mean below.

In the cell link box enter NavMenu!$D$9. It is important to specify the NavMenu worksheet since you'll use this on many pages and it must only modify the cells in the NavMenu worksheet, not the current sheet. Now you should have a combo box that contains all the menu items and posts the index in cell D9.

Step 5: Cell D10 is the name of the menu item. To do this we use the offset command. In Cell D10 enter =OFFSET(B8,D9+2,0,,) . This takes B8 as the reference cell, counts value(D9)+2 rows down, and displays whatever is in that cell. In our case, it's the menu item name.

Step 6: In Cell D8 enter =COUNTA(B11:B18). This is not shown in the picture above because i just added it. It counts the total number of non blank cells in the range. I include B18 because when I insert new item, the range will automatically adjust. New items are inserted in the menu by insert (shift cells down) at B18.

Step 7: Open up the Developer menu (search google if it's not already displayed in your ribbon) and visual basic. Create a new module for the menu and paste:

Sub TournamentMenu()

Dim PageName As String, GroupName As String, SheetName As String

Dim ItemNum As Integer

ItemNum = ['NavMenu'!D8] 'number of menu items

PageName = ['NavMenu'!D10] 'page name

GroupName = ['NavMenu'!B8] 'group name

If PageName = "0" Then 'condition to exit if blank

Exit Sub

End If

SheetName = GroupName & " " & PageName 'join the groupname and pagename

Sheets(SheetName).Select 'navigate to the sheet

['NavMenu'!D9] = ItemNum + 1 'reset the menu to the blank entry

End Sub

Step 8: Right click on the combobox and click assign macro. Select TournamentMenu and you're done!

Repeat for different menus. Modify the cell locations for ItemNum PageName and GroupName for each menu.

Generating Unique Random Numbers in VBA Excel

2011-03-10T14:00:00.000-08:00

Problem: Generate a list of unique random integers between two bounds in VBA and Excel

Solution: I started with the code from here. The original code output the random numbers as a string residing within the same cell. Needed each random number to appear in its own cell. The context was to randomize seating for a poker game, hence the variable names players

I pull the input data from the excel cells directly so adjust your code to match the cells you want to input:

Top - upper bound

Bottom - lower bound

Players - cell range of players (The range of cells next to the list of player where each player name is located in a column of cells)

Execute code on button click

Sub Seating()

Dim iArr As Variant

Dim i As Integer

Dim r As Integer

Dim Top As Integer

Dim Bottom As Integer

Dim temp As Integer

Dim Players As Range, Cell As Object

Set Players = Range("W10:W25")

Top = [V7]

Bottom = 1

Application.Volatile

ReDim iArr(Bottom To Top)

For i = Bottom To Top

iArr(i) = i

Next i

For i = Top To Bottom + 1 Step -1

r = Int(Rnd() * (i - Bottom + 1)) + Bottom

temp = iArr(r)

iArr(r) = iArr(i)

iArr(i) = temp

Next i

j = 1

For Each Cell In Players

Cell = iArr(j)

j = j + 1

If j > Top Then

Exit For

End If

End Sub

Modifying the Heading Style for Bibliographies in Latex

2011-03-08T13:39:00.000-08:00

http://www.latex-community.org/forum/viewtopic.php?f=5&t=4089

The best way

\renewcommand\bibsection{\subsection{\refname}}

where "\subsection{}" says what type of heading to use (change it to \section or \chapter or whatever) and \refname gives it the name Reference but can be changed to whatever you want.

For UTK dissertations they require a section page to separate the actual bibliography where the TOC references the essentially blank page. So stupid. It's really easy in word but if you were using word, you wouldn't be reading this. My code looks like

\addToTOC{Bibliography}

\thispagestyle{plain}

\chapter*{Bibliography}

\newpage

\renewcommand\bibsection{}

\bibliographystyle{ieeetr}

\bibliography{references/References}

Rotating a Single Page in Latex

2011-03-08T12:32:00.001-08:00

It seems common (and absolutely, positively, completely asinine) that dissertations and theses require that rotated figures appear on rotated pages. I mean that the page (headers, footers, page number, etc) are rotated with the figure so that the whole page is in landscape and not just the figure. Apparently they (whoever "they" are...I suspect a primitive race of subhumans that haven't developed advanced logic...or liberal arts majors) think that in this digital world landscape figures should appear vertically in the electronic document by rotating the whole page so all the text appears vertically. This is dumb because it forces a different formatting of fancy header/footer positions but most importantly because the hard copy version (THE REAL DOCUMENT THAT SOMEONE HOLDS IN THEIR ACTUAL HAND) has to be rotated to read the running header. That is not how a document flows even if the document has to be rotated to read an ELEMENT (table, fig., etc.). The ELEMENT should not dictate the PAGE SETUP. It should be the other way around! Documents with running headers use that information to track the location in the document even if the table is not useful. Hence, the document should not be rotated to continue tracking the location. Anyway, since my tiny brain can't compete with that of the clearly more intelligent liberal arts mind, I have to format MY dissertation in this stupid way!

Whew...breathe Josh...

Problem: My document contains a long table that fits in landscape but not portrait view. The entire page needs to be rotated including headers and footers.

Solution: What solution!? They can shove it up their... Sorry... got excited again. Thanks to Alan Munn we have a reasonably sophisticated answer. I saw some other ones (simpler ones) that worked too, but this seems to be the most correct within the context of programming in Latex. He posted it here. I took what he had and edited it for the UTK Latex Template. This is the exact code for one of my long tables

\documentclass[letterpaper,12pt]{report}

\usepackage[vcentering,dvips]{geometry} % help with margins

\usepackage[absolute]{textpos}

\usepackage{booktabs}

\usepackage{pdflscape}

\geometry{

letterpaper,

left=1.0in,

top=1.5in,

right=1.0in,

bottom=1.5in,

paperheight=11in,

paperwidth=8.5in,

dvips,

pdftex

}

\pagestyle{fancy}

\fancyhead{} % clear all header fields

%\fancyhead[LE,RO]{\slshape \rightmark}

%\fancyhead[LO,RE]{\slshape \leftmark}

\fancyfoot{} % clear all footer fields

\fancyfoot[LE,RO]{\thepage}

\renewcommand{\headrulewidth}{0.4pt}

\renewcommand{\footrulewidth}{0.4pt}

\fancypagestyle{lscape}{%

\fancyhf{} % clear all header and footer fields

\fancyfoot[LE]{% Left field on EVEN pages

\begin{textblock}{20}(1,5){\rotatebox{90}{}}\end{textblock} % This is for running header info \leftmark or \rightmark

\begin{textblock}{1}(13,10.5){\rotatebox{90}{\thepage}}\end{textblock}}

\fancyfoot[LO] {% Left field on ODD pages

\begin{textblock}{1}(18.75,3.5){\rotatebox{90}{\thepage}}\end{textblock} % this positions the page number

\begin{textblock}{20}(1,13.25){\rotatebox{90}{}}\end{textblock}}% This is for running header info \leftmark or \rightmark

\setlength{\TPHorizModule}{1cm}

\setlength{\TPVertModule}{1cm}

\renewcommand{\headrulewidth}{0.0pt}

\renewcommand{\footrulewidth}{0.0pt}

}

\begin{document}

\newpage

\thispagestyle{lscape}

\begin{landscape}

% EXPERIMENTAL MANTLE LOCATIONS

%===============================================================================================

%\begin{sidewaystable}

\begin{table}

\addtolength{\tabcolsep}{-2.0pt}

\centering % ensure it centers on the page. modify this if centering is not wanted

\caption{Comparison of experimental and computational mantle locations with the complex-lamellar (\text{CL}) and Beltramian (\text{BEL}) models}

\label{Ch:Vortex Sec:MULT Comparison Experimental mantle locations} % this is the label for cross-referencing

%\begin{tabular*}{6.5in}{@{\extracolsep{\fill}} lllllccclcc }

\begin{tabular}{lllllccclcc}

\toprule

& & & & \multicolumn{3}{c}{Experimental} & & \multicolumn{3}{c}{Computational} \\ \cmidrule(){5-7} \cmidrule(){9-11}

$n$ & $\beta_{2,n}^{({\rm CL})}$ & $\beta_{2,n}^{({\rm BEL})}$ & & $\beta_{2,n}^{({\rm EXP})}$ & $\left |\beta_{2,n}^{({\rm CL})}-\beta_{2,n}^{({\rm EXP})} \right |$ & $\left |\beta_{2,n}^{({\rm BEL})}-\beta_{2,n}^{({\rm EXP})} \right |$ & & $\beta_{2,n}^{({\rm CFD})}$ & $\left |\beta_{2,n}^{({\rm CL})}-\beta_{2,n}^{({\rm CFD})} \right |$ & $\left |\beta_{2,n}^{({\rm BEL})}-\beta_{2,n}^{({\rm CFD})} \right | $ \\

\hline

\\[-2.0ex]

0 & 0.408 & 0.236 & & 0.296 & 0.112 & 0.060 & & 0.305 & 0.103 & 0.069 \\

1 & 0.707 & 0.543 & & 0.594 & 0.113 & 0.051 & & 0.385 & 0.322 & 0.158 \\

2 & 0.913 & 0.851 & & 0.803 & 0.110 & 0.048 & & 0.787 & 0.126 & 0.064 \\

\bottomrule

\end{tabular}

\end{table}

%===============================================================================================

\end{landscape}

\end{document}

My Latex Custom Definitions

2011-03-08T08:31:00.000-08:00

This set goes either in the preamble of my latex documents or in the class file. It is convenient to switch between publication formats by just changing the definition here rather than throughout the document.

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% REFERENCING Commands

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% EQUATIONS

\renewcommand{\eqref}[1]{Eq.~(\ref{#1})} % Eq. (1)

\newcommand{\seqref}[1]{Equation (\ref{#1})} % Equation at the start of a sentence

% MULTIPLE EQUATIONS: CAPTURES THE BEGINNING & ENDING EQN NUMBERS

\newcommand{\eqsref}[2]{Eqs.~(\ref{#1}--\ref{#2})}% Eqs. (1-3) in text: use \eqsref{eq1}{eq3}

\newcommand{\seqsref}[2]{Equations (\ref{#1}--\ref{#2})} % Equations (1-3) at the start of a sentence

% PARTS OF EQUATIONS WITH LABELS / Eq. (1a)

\newcommand{\eqrefa}[2]{Eq.~(\ref{#1}{#2})}% Eq. (1a) : with a suffix in text

\newcommand{\eqsrefa}[2]{Equation (\ref{#1}{#2})}% Equation (1a): with suffix at sentence beginning

% FIGURES

\newcommand{\figref}[1]{Fig.~\ref{#1}} % Fig. 1

\newcommand{\sfigref}[1]{Figure \ref{#1}} % Figure 1 at the start of a sentence

% MULTIPLE FIGURES

\newcommand{\figsref}[2]{Figs.~\ref{#1}--\ref{#2}} % Figs. 1-3 in text: use \figsref{fig1}{fig3}

\newcommand{\sfigsref}[2]{Figures \ref{#1}--\ref{#2}} % Figures 1-3 at the start of a sentence

% PARTS OF FIGURES WITH LABELS

\newcommand{\figrefa}[2]{Fig.~\ref{#1}{#2}} % Fig. 1a -may use \ref{#1}{\it{#2}}} for a different style

\newcommand{\sfigrefa}[2]{Figure \ref{#1}{#2}} % Figures 1a at the start of a sentence

% TABLES

\newcommand{\tblref}[1]{Table \ref{#1}} % TABLES

\newcommand{\stblref}[1]{Table \ref{#1}}

\newcommand{\tblsref}[2]{Tables \ref{#1}--\ref{#2}} % TABLES in text: use \tblsref{tbl1}{tbl3}

% SECTIONS

\newcommand{\secref}[1]{Sec.~\ref{#1}}

\newcommand{\secsref}[2]{Secs.~\ref{#1}--\ref{#2}}

\newcommand{\ssecref}[1]{Section \ref{#1}}

\newcommand{\ssecsref}[2]{Sections \ref{#1}--\ref{#2}}

\newcommand{\algref}[1]{Alg.~\ref{#1}}

\newcommand{\algrefs}[2]{Algs.~\ref{#1}--\ref{#2}}

\newcommand{\salgref}[1]{Algorithm \ref{#1}}

\newcommand{\salgsref}[2]{Algorithms \ref{#1}--\ref{#2}}

\newcommand{\appref}[1]{App.~\ref{#1}}

\newcommand{\appsref}[2]{Apps.~\ref{#1}--\ref{#2}}

\newcommand{\sappref}[1]{Appendix \ref{#1}}

\newcommand{\sappsref}[2]{Appendices \ref{#1}--\ref{#2}}

\newcommand{\chapref}[1]{Ch.~\ref{#1}}

\newcommand{\chapssref}[2]{Chs.~\ref{#1}--\ref{#2}}

\newcommand{\schapref}[1]{Chapter \ref{#1}}

\newcommand{\schapsref}[2]{Chapters \ref{#1}--\ref{#2}}

%----------------------------------------------------------------------------------------------------%

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% VECTORS

\newcommand{\bvec}{\ensuremath{\boldsymbol}} % Change \boldsymbol to \mathbf for non-italic vectors

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\newcommand{\ubf}{\ensuremath{\bvec{u}}}

\newcommand{\ubfbar}{\ensuremath{\bar{\ubf}}}

\newcommand{\ubftilde}{\ensuremath{\tilde{\ubf}}}

\newcommand\nbf{\ensuremath{\bvec{n}}}

\newcommand \e {{{\bvec{e}}}}

\newcommand \ex {{\e}_x}

\newcommand \ey {{\e}_y}

\newcommand \ez {{\e}_z}

\newcommand \er {{\e}_r}

\newcommand \et {{\e}_{\theta}}

\newcommand{\Omegabf}{\ensuremath{\bvec{\it{\Omega}}}}

\newcommand{\omegabf}{\ensuremath{\bvec{\it{\omega}}}}

% VARIABLES

\newcommand \rhobar{\ensuremath{\bar{\rho}}}

\newcommand \rhotilde{\ensuremath{\tilde{\rho}}}

\newcommand \thalf{\ensuremath{{\textstyle\frac{1}{2}}}}

\newcommand \etal{\mbox{\textit{et al.}}}

\newcommand \etc{etc.\ }

\newcommand \eg{e.g.\ }

\newcommand \V {\ensuremath{\mathcal{V}}} % for control volume

\newcommand \Vcal {\ensuremath{\mathcal{V}}} % for control volume

\newcommand \G {\ensuremath{\mathcal{G}}} % for script function G

\newcommand \F {\ensuremath{\mathcal{F}}} % for script function F

\newcommand \Gcal {\ensuremath{\mathcal{G}}} % for script function G

\newcommand \Fcal {\ensuremath{\mathcal{F}}} % for script function F

\newcommand \Sf {\ensuremath{\mathcal{S}}} % for control surface

\newcommand \Scal {\ensuremath{\mathcal{S}}} % for control surface

% INTEGRATION

\newcommand \dSf {\ensuremath{\, {\rm d} \mathcal{S}}} % surface

\newcommand \dV {\ensuremath{\, {\rm d} \mathcal{V}}}

\newcommand \dr {\ensuremath{\,{\rm d}r}}

\newcommand \dtheta {\ensuremath{\,{\rm d}\theta}}

\newcommand \dphi {\ensuremath{\,{\rm d}\phi}}

\newcommand \dx {\ensuremath{\,{\rm d}x}}

\newcommand \dy {\ensuremath{\,{\rm d}y}}

\newcommand \dz {\ensuremath{\,{\rm d}z}}

\newcommand \dt {\ensuremath{\,{\rm d}t}}

\newcommand \dd {\ensuremath{\,{\rm d}}}

Placing Multiple Figures on a Beamer Slide

2011-03-03T12:28:00.001-08:00

Beamer is pretty interesting. Since I prefer Latex for my documents it makes sense to try it. I'm not 100% sold, but it's pretty damn aesthetic. What I don't like, is that it doesn't seem to take the same syntax as the actual Latex document. For instance, multiple figures with individual captions don't seem to work since the subfloat package seems to crash it. The way around is to have syntax like

\begin{figure}

\centering

\includegraphics[width=.45\textwidth]{../figures/Ch_Vortex_Flows/CL_Viscous_Ur}

\includegraphics[width=.45\textwidth]{../figures/Ch_Vortex_Flows/CL_Viscous_Uq}\\

\includegraphics[width=.45\textwidth]{../figures/Ch_Vortex_Flows/CL_Viscous_Uz}

\includegraphics[width=.45\textwidth]{../figures/Ch_Vortex_Flows/CL_Viscous_P}

\caption{The viscous complex-lamellar bidirectional vortex with $\kappa=0.103$ and $z=0.3$.}

\end{figure}

It's pretty compact compared to all the stuff for the subfigures in the draft. As far as I can tell, if you want multiple captions, then you have to put the graphics in individual \begin{figure} blocks but then the vertical spacing won't fit well.

Sage Server Through VMware

2011-03-01T08:35:00.000-08:00

Sage is a open source frontend and server that I use to replace Mathematica in particular, but will also serve as a frontend for Mathematica, Matlab, Latex, Python (it's foundation), and more. It is almost powerful enough to completely replace Mathematica as an analytic mathematics software. It's really close. It surpasses Mathematica in the ease and speed in which you can do real coding for numerical mathematics as well. It uses Python, numpy and scipy so all the standard Matlab stuff is included. It spits out really nice figures too. The server side is one of the best parts! Documentation is kinda poor to get the install right but once you do, it will automatically store accounts and log in information for multiple simultaneous users. It's browser based, so you never have to install anything on your personal machine and it just so convenient!

Enough about that! The biggest problem - poor documentation for the novice since it doesn't have a Windows version. It has a virtual machine version and ok documentation to install with Windows as the host but screw that! If you're a research group or company, I would forgo all the VMware stuff I'm including here and just build a dedicated Linux server. You'll get more performance, less set up, and you can use it to run all kinds of other code scripts so each client machine doesn't have to have all that garbage installed.

I had help setting my instance from this document. I did things a little differently and that's why i'm posting about it now.

Install VMware. Get VMware. We had a copy of VMware Workstation that was unused so I know it works. Virtual Box probably works too.
Install Linux on VMware. I installed Ubuntu x64. Overall, Ubuntu is the least foreign to my simple Windows using mind;). File-New-Virtual Machine-Install from iso-follow the onscreen directions
If this is a VMware install, I assume this is a test bed and not a full implementation. If this is true, select the amount of RAM and Disk Space you want to allocate in the OS settings through VMware. The Network Adapter should be set to Bridged: Connect directly to the physical network. The checkbox "Replicate physical ..." is unchecked. This step is important for assigning a unique IP address to the Virtual Machine so it can be accessed directly and not just through the host.
Install an ssh server. Sage is a command line software (so is Mathematica, Matlab, Python...etc) with a frontend attached. Command line access through ssh can be helpful if you want to remotely run code...any code...not just sage...so I recommend doing this regardless of what you plan on putting on your server. Do this with sudo apt-get install openssh-server
Download Sage from sagemath.org. Make sure you get the right version for your operating system.
Install. I never remember how to install anything in Linux...I can't understand why Linux operating system designers make it so damn complicated. I think the correct install command for sage is sudo ./ but it might simply be sudo ./sage ...I don't really remember but I think it actually just extracts somewhere and is done so you might want to make sure the install file is already in the folder you want the final install in...I just can't remember. It might actually have a debain package that "just works" when you double click it.
Test to make sure it installed. open terminal and type the word "sage" you should get a sage: prompt. If you type notebook() a web browser should pop up with the generic server index page. More specifically type notebook(address='',port=8000,accounts=True) to start a server on that ip with that port and user accounts enabled.
To really make this a server, it is convenient to set up some symbolic links. Enter (from the document linked above)

ln -s sage

cd /usr/local/bin

sudo ln -s /sage sage

sudo ln -s /local/bin/sage-python sage-python

Here, is the full path to the Sage directory, something

like:

/home//sage-3.2.1-etc...

To test the links, go to your home directory and type:

sage

This should start Sage. Then enter:

sage-python

This will start the version of Python that comes with Sage.

Now we want a script to start the server with all the options already set. I'm not sure if it has to be a Python script but Python plays nice with Linux and that's a good thing. In the home directory make a file called notebook.py with the following commands inside

from sage.all import*

notebook(address='XXX.XXX.XX.XXX',port=8000,accounts=True)

all the Xs are for the IP address of the virtual machine. make sure the apostrophes are included. accounts=True allow for private user accounts to be created on the server. When you run this script terminal will give an error about being crazy if you don't have secure=True but True seems to disallow access over the internet.

Start the server with sage-python notebook.py
Make a launcher from the desktop. I think there is a way to have the server start with the machine but I'm to dumb to use Linux. My small brain can only hand simple Windows and the super convenient Startup folder. If you're like me, right click on the desktop and create launcher. It's either type: application or application in terminal. I don't remember. and the command: sage-python notebook.py
THIS STEP PROBABLY ISN'T NEEDED FOR BRIDGED NETWORK ADAPTERS. From Windows go to Allow program through firewall and open port 8000 in UDP and TDP.
The server can now be accessed through any web browser by typing http://XXX.XXX.XX.XXX:8000 or by pointing any ssh client to the same address

Sage is sweet! I can access and share my personal notebooks through any browser - including my phone. You can make really professional documents and reports directly in a web browser if you know a little html programming too. I hope this project continues. It is one of the most remarkable examples of how open source software can actually be very good!

Review: Gemaco Sextant Playing Cards

2011-02-09T07:36:00.000-08:00

Gemaco Sextant are the product of a "reformulation" by the Gemaco company. What this really means is they are a different and probably lower standard when compared to their front runner, the Gemaco Star. They do initially appear to be like a discount plastic playing card material but play much better. This doesn't mean these are bad cards. They are slightly less expensive then the Stars but still in the same price range as all the rest. The first thing you'll notice compared to the rest of the cards reviewed here is that they are very shiny. Where most of the brands have a textured surface these are completely smooth. At first expected them to be slippery but in fact they are not and probably less slippery than the rest. The reason is that the texture of the other cards is purposed to create an air layer between the card and the playing surface or other cards. The air layer allows the cards to slide easily on felt as well as shuffle smoothly. These cards do not generate this layer and, as such, don't slide across the felt well. This isn't a bad thing. They do slide fine and certainly well enough to facilitate a smooth game. They just don't slide AS MUCH. This could be an advantage for players who have trouble controlling the pitch. The also shuffle well. If the Stars can be compared to Kem then these are compared to Copag. These are much thinner - part of why they are easy to shuffle - while retaining good elasticity and snap. I think overall, my regular group really liked how these cards played.

Appearance wise, they are similar to the Stars with the same print of numbers, pips, and paint. The backs are even similar. Also consistent with the stars is the use of brilliant colors. The reds are SO red and the blues are very blue. I really love the look of Gemaco. The ace of spades is too large but very pretty.

Unfortunately, these cards do mark. After a couple hours of playing one card had a definite crease where a fingernail pushed down onto it while peaking at the faces. This was heartbreaking since I do like these cards but one deck was ruined the first night! I was the guilty player but I am certain that I did nothing out of the ordinary when peaking and I've been playing live poker for quite some time now. They are a good card but definitely shy of the Gemaco Stars but seem to be a very well received choice for my group and maybe yours.

Python: The Division Sign

2011-02-06T08:17:00.000-08:00

I've been wrong all these years! I'd always thought 3/2=1.5 . Apparently, it's actually 3/2=1. Much like pi=3 to the Alabama state legislature instead of pi=3.14159... (see the joke).

Clearly this is "wrong" (at least for scientific computing and arguably everywhere else) but here's why this happens. It's the difference in how the data types are defined - the output depends on the input types so

3/2=1 because 3 and 2 are int (integer) type so the output is a truncated integer. This is "floor division" and is how C does it too. However,

3.0/2=1.5 because 3.0 is a float so the output is a float even though 2 is still typed in as an int. I think it's actually a float now too but i'm not sure.

Starting with Python 2.2 a new division operator is introduced. Now

3/2=1.5

3//2=1

The "//" is now the floor division and the single "/" will be true division. I think Python 3 will have this completely integrated but all us 2.x users we're still SOL unless we want to rework all our codes.

Solution: import a module to overwrite the division sign with the upcoming (and computationally relevant) definition. At the very top include the statement

from __future__ import division

now the "/" will always be true division.

This module probably has a bunch of other important stuff. Since i'm learning python right before these all become standards is there a way to import ALL the future commands so I don't have to relearn anything when 3.x if viable for me? Something like from __future__ import * would be nice!

Python: Matplotlib for 64bit

2011-02-06T08:06:00.000-08:00

If you have 64bit python installed...I used the Enthought package to do it...Matplotlib probably won't actually work because it's not available for 64bit in general. So when you try to plot something you get a huge error output. If you look closely, the only real error is that it can't find a font in the "show()" command because it's looking in the (x86) directory. The brute force fix - find the correct subdirectory in the 64 bit install directory and copy it into the equivalent 32 bit install directory. In short:

Copy

C:\Program Files\Python26\Lib\site-packages\matplotlib

Create/Paste

C:\Program Files (x86)\Python26\Lib\site-packages\matplotlib

duh as in done!

I have a feeling all that needs to be done is add/change the path so it can find the right folders anyway. If anyone reads this and can confirm, please comment. Thanks!

Python: Tips for Matlab Users: Standardizing Syntax

2011-02-06T06:46:00.001-08:00

Remember how in Matlab you have to call libraries and denote commands with prefixes referring to the library they are hosted in? You don't? Because you DIDN'T:). You do in python though. This is really because python is not originally a scientific computing software but it is supposed to allow for addons, packages, and all kinds of extension including scientific computing. Interestingly enough, this is also the way to call functions from other py files (that's vaguely familiar in Matlab).

But python is a similar to C with respect to loading headers and similar to Matlab in the implementation. Like C we have to import our "modules" or "libraries" (same thing) or specific "functions." Before an example read this article about the "from" and "import" commands.

For our an example, lets solve a complex generalized eigenvalue problem, calculate the execution time and plot the spectrum. We need scipy (numpy can only do single matrix eigenvalue problems) and a time function. More specifically, we need the eig() function inside the linalg library in scipy.

First the correct and safe way:

Here we want to import the eig() function from the scipy.linalg library, then prefix all scipy commands with "sc" in case we want to use modules with conflicting function names. This requires two import statements. One for the special function eig and one to denote the scipy commands.

Next we want the plot library available so we import the matplotlib.pyplot with the prefix plt. The pyplot is the special collection of functions inside matplotlib to make nice plots - just like eig was in the linalg library in scipy. Finally we want to import the time function

Now we define the matrix size and generate random NxN complex matrices A and B. Start the clock. Compute the eigenvalues and eigenvectors (L,V). Generate the figure and define the active object number. Scatter plot the spectrum. Force the plot to draw. Display the calculation time. All this is shown below.

################################################

import scipy as sc

from scipy.linalg import eig

import matplotlib.pyplot as plt

import time

N=50

A=sc.random.random((N,N))+sc.random.random((N,N))*1j

B=sc.random.random((N,N))+sc.random.random((N,N))*1j

tic = time.time()

(L,V) = sc.linalg.eig(A,B)

toc = time.time()

print toc-tic, " has passed"

plt.figure(1)

plt.scatter(sc.real(L),sc.imag(L))

plt.show()

################################################

Notice how i use the sc prefix on every scipy command. This might be important if i'm also using built in sympy commands with the same function names...but i'm not. Let's do this again. This time no prefixes.

################################################

from scipy import *

from scipy.linalg import *

from matplotlib.pyplot import *

from time import *

N=50

A=random.random((N,N))+random.random((N,N))*1j

B=random.random((N,N))+random.random((N,N))*1j

tic = time()

(L,V) = eig(A,B)

toc = time()

print toc-tic, " has passed"

figure(1)

scatter(real(L),imag(L))

show()

################################################

The from import * commands give access to all the functions in the library...not just the eig function... and with no prefixes. The random still has a prefix and i'm not sure why or how to make it go away but I started learning python about 48 hours ago so oh well. Either way, this is similar syntax to matlab so python should be a viable alternative!

FYI run times for both are about the same for the eig command with both running 64 bit versions.

Python: Tips for Matlab Users: Getting Started with Python

2011-02-05T20:53:00.001-08:00

Engineers are not computer scientists and as such Matlab is perfect for us guys that don't want to screw around with nonsensical syntax. Computer science guys are so damn pretentious about their programing languages i'm sure some of them will read this and get a little red in the face. Let's face it though; our job is NOT to write elegant codes. Our job is to solve practical engineering and applied mathematics problems. With Matlab, we don't have to spend an hour trying to find where the damn missing bracket is supposed to go (i hate you most c and c++ lol). It's clean, simple, powerful, and it just works. The trade off is speed...sometimes...

Enter python and more specifically numpy, scipy, ipython, and matplotlib.

But first - why python? Beats me...but it was recommended to me. That's not 100% true though. Mostly it's simplicity of syntax compared to c and built in fortran wrappers for those awesome Lapack, Linpack, Blas...(idk if they are all fortran but whatever...they are all there). In fact with a little set up, the python commands can be reduced to nearly the same as Matlab.

Problem #1: As of early 2011 python is experiencing a large overhaul from v2.x to v3.x but numpy, scipy, and matplotlib aren't ported yet.

Problem #2: The sponsored numpy download installs over python 2.6 - python 2.7 is the currently available release. Idk where 2.6 is hosted.

Problem #3: Scipy needs numpy. Scipy installs over python 2.7. Numpy installs over 2.6. Scipy can't be installed since numpy can't be installed.

I could probably figure out #2 and 3 but back to the original statement - why should I have to. You developers are the computer science guys. If that's the best you can do...then I hope you never get hired to develop something for the general public!

Solution: Enthought. It packages it all together with one install so you don't have to screw around. Everything mentioned here is for fundamental scientific computing and visualization so specialized packages may or may not be available in python. Enthought has 32 and 64 bit versions and if you have a 64 bit operating system definitely use the 64bit version or else you won't have the same speed optimization that Matlab offers. In reality Enthought probably doesn't offer anything unique but it simplifies and consolidates the install process. It's not free (academic licenses are) but it's a hell of a lot cheaper than Matlab!

Review: Gemaco Star Playing Cards

2011-02-03T15:28:00.000-08:00

Gemaco Stars are another set of 100% plastic playing cards. I'm not sure what kind of plastic these are though. The depictions to the left are accurate. These cards are about the same thickness as a deck of Kems but don't feel the same. They are very unique in that they feel "soft" to the touch. Not soft like malleable but soft like silky. They are not slippery though. They are comfortably flexible but stiffer than the Copags. More on par with Kem. They riffle very well and slide together well also. All in all, they are VERY playable. If you like Copag but want a heavier card, then forget Kem and buy these.

The colors are very bright on the backs but the facecards rely heavily on black with dark thick lines. The eyes are also very large on the faces but beautifully drawn. Unlike many cards, all the kings, queens, and jacks have unique faces. I personally find that very appealing. The ace of spades is decent but WAY too big! It could use a serious reduction in size if it wants to be the prettiest card in the deck.

I bought a "used - very good" deck from amazon and was disappointed to find the graphics were misprinted. By misprinted i mean not centered. In all truthfulness, it's probably just my decks and that's why i got them cheap.

That aside, I Love these cards after playing only once with them. If they hold up they easily rival Copag as my favorites.

Review: Bicycle Prestige Playing Cards

2011-02-03T15:11:00.000-08:00

Everybody knows Bicycle for their paper playing cards but few realize they have a competitive plastic playing card too. The Prestige cards were originally released as a limited time offer but apparently were successful enough for continued production. These were my first deck of "real" plastic playing cards and they are interesting. They boast a "paperlike feel" and I guess they do now after plenty of use. It could be the Dura-Flex plastic (I suspect is PVC although the original releases were prone to "acetate" smells? Since most people are morons, I'm not sure. Maybe the moron is me this time. Regardless, mine didn't smell and don't feel like cellulose acetate.) or it could be skin oil collecting on the surface but either way, they do feel most like paper than any other card I own. To be clear, i mean "like paper" in how they feel when shuffling. Otherwise, they are thick and fairly stiff. Less stiff than the Da Vinci cards but more so than the Kems

The color is very pretty for both the red and blue. The picture is pretty accurate for both. The backs are beautiful and look very good spread out while washing the deck. The face cards are pleasing and are no different than those printed on the paper cards. The ace of spades is one of my favorites. It possesses enough detail to make it stand out but not so much as to look cluttered. Also, it is sized appropriately unlike the Copag (too small), Da Vinci, or Gemaco (too big).

The problem with these is that they float easily and some people will have a problem getting them to stop moving when they set them down. The solution is easy - give a quick tap to the top card to push all the air out. A second problem i've noticed is the ink is flaking slightly after about 6 months of use.

Either way, these are decent. Not my favorites but not bad either. Definitely good enough to play with on a regular basis. Maybe the best part, you can buy them at Walmart.

Review: Modiano Da Vinci Playing Cards

2011-02-03T14:59:00.000-08:00

I picked up a set of Modiano Da Vinci playing cards after being well recommended against Kem and Copag on a youtube review. They are PVC (i think) plastic cards but much thicker than Copag. They are about as thick as a deck of Kem. Shuffling is ok with these, but not as good as Kem, definitely not as good as Copag and maybe not as good as Bicycle Prestige. They slide together fairly well but are very stiff and don't riffle as smoothly as other cards. Pitching is easy given their weight and peaking is satisfactory given their stiffness.

The red is close to the color depicted but the blue is much darker. Many of the amazon pictures show a very bright blue but this is far from the almost navy back. The detail in both the back and face cards is very precise and will not disappoint. The font for the numbers is a very elegant, thin font with tapered edges while the face cards are all bold block like font similar to the Kem print. This mismatch detracts from the overall appearance in my opinion. The pips are rounder and fatter in general and are quite pleasing. The ace of spades is elaborate and very nice, but probably not my favorite.

The biggest problem with these is the edges of the cards. They are extremely sharp and while not able to cut skin, after a whole evening of dealing, my hands were raw like I'd been rubbing sandpaper. I would not recommend these personally, but I know two (including the youtube reviewer) people that really like them!

Review: Kem Arrow Playing Cards

2011-02-03T14:46:00.000-08:00

Kem is probably the most trusted name in playing cards. These are 100% cellulose acetate playing cards. That means nothing to me except that you can expect these to feel like a softer, more malleable plastic than a PVC card. I enjoy these cards very much. They have excellent flexibility an snap. They may feel slightly more slippery than Copag cards, but not so much that people should have problems handling them. They are about as thick as a deck of bicycle paper cards and maybe a little thicker. They shuffle well, but not as well as a Copag but probably pitch slightly better being heavier. They do not slide together as well as one might hope after playing with Copag but they are still very playable.

The colors are deep blue and blood red. The facecards are also deep in color and rich in detail. The ace of spades is likely the most famous rendition in production and very appealing. The font is a traditional playing card typeset and easy to read while remaining appealing to the eye.

These cards will warp. As they heat up over the course of a tournament they will become very malleable and retain the shape of the dealers hand. This is not to say they will be markable but they probably won't lay completely flat at the end of the night. They also have a history of warping outside of play if not pressed in between sessions. To remedy this either press them till the next game or place a piece of cardboard on front and back and wrap them with rubber bands. They will be ready for the next game.

Review: Copag 1546 Playing Cards

2011-02-03T14:18:00.000-08:00

Copag is one of the most popular 100% plastic playing card manufacturers and with good cause. These cards are PVC plastic - as are many plastic playing cards - but the defining characteristics of these particular cards makes them unique. They are thin...very thin. If your preference is toward cards about the same thickness as a standard paper deck then you might not like these. Even so, I would strongly recommend trying them. The thin profile and distinct surface texture makes professional shuffling exceptionally easy. They riffle better than any card I've played with and slide together very smoothly. Pitching the cards, although not hard, is a little different since they are lighter. Peaking at the corner is exceptionally easy. For holdem, this is nice. For draw or 7 stud a thicker card might play better but I love these cards all around.

The color is decent. The blue is a nice soft tone but the red is a little orange. The scroll work on the back is second to none but the face cards aren't exceptionally detailed. The print is like that of a typewriter and isn't the most aesthetic either. Lastly, the ace of spades is the most bland rendition I've ever seen. Of course, none of these characteristics detracts from playablity. In fact, the playablity surpasses the aesthetics so much that I feel guilty saying anything bad about their appearance.

They do have a significant defect. The ink will bleed in humid environments. In fact, running a humidifier near them is enough. The blue will run first but the red will follow. They do not however, warp.

Even this cannot shake them as one of my favorite two cards for shear playablity alone.

Transparent Backgrounds in OriginLabs Plots

2010-12-01T06:54:00.001-08:00

I was trying to overlay a layer with constant contour lines over a color fill contour plot with different contour lines to make a comparison between the two.

Adding the contour lines to the existing plot doesnt work because...and to remain consistent with the complete disregard for user friendliness...it "kinda" links the magnitude of the constant lines with the color map for the color fill plot. By "kinda" I mean it's a mess and is a terrible software that offers a lot of options and I can't get away from it AHHHHHH.

Solution: use colored markers...err...add another layer, but of course that doesn't work straight forwardly either. Even if all the layer management options for fill and background are "none" it isn't so you won't see the first layer plot. Solutions...light your computer on fire...err...open up format>layer>Size/Speed>Graphic Image Caching and select something else. "None" worked for me but this link suggests that there are other options.

Outlook 2010 Login Fail for POP3

2010-07-19T12:40:00.000-07:00

Outlook 2010 is pretty decent but it is following in the path of many other softwares that actually lose functionality as it matures. Outlook 2007 had a great feature where you could just put a username, password and email address in their respective boxes and it would automatically set it up. This has failed miserably in 2010.

Trying to set up a POP3 email repeatedly asks for a password and returns something about the server rejected the login info. If you copy your old 2007 settings exactly, it still fails.

Solution (for me): Click on repair and more settings. Under the outgoing server tab uncheck the first box so the options gray out. Also, don't require logon Secure Password Authentication.

Strange Lines in eps File Print Outs

2010-07-11T17:28:00.001-07:00

I can't imagine anyone else needing this post but maybe it'll save some poor soul hours of screwing around too.

I had plots generated in Origin that contained the function 1/r. I put them into the latex document and converted to pdf. Everything looked great in pdf form. When I printed it, vertical lines (slightly slanted) appeared that went from the top edge of the graph to the function. If I opened the eps in adobe and printed it, the lines shifted but were still present. If I printed from Origin they looked fine. If I printed to PDF from Origin, they were fine. If I printed to ps and converted to pdf they appeared again. I knew it wasn't a erroneous data point since they didn't show up anywhere else and shifted position depending on how I converted them.

Solution (sort of):
Use less data points. I was using 1000. The line disappeared when I used 500.

I have NO idea where or what the problem was but there you go.

UPDATE* This problem persists and is total BS!!! The solution above didn't work for another example. What i noticed was it only occurs if there is a lot of data points outside of the plot range. I thought, well maybe for some reason the points outside of the plot area are connecting with the points inside. I though this because only the data range containing points outside the graph had a problem.

Solution! Don't include the full plot range. If you don't include the largest datapoint the stray line changes position, if you dont include the largest 2, it changes again. At some point it will disappear. I don't know how many you need to eliminate but one to a few. The easiest way to do this is to go to the graph, right click, plot setup, click on the offending graph data, click the button that appears next to the range (has 3 dots on it) and change the starting row to not include the first few data points. This way, if you want to rescale the axis to include more data it still remains. The other option is deleting it manually. This is much quicker but not advised.

Automatic Positioning of Superscript References in Latex

2010-06-30T11:38:00.000-07:00

Some publications require superscript references and some numeric, inline references. The first should appear like

blah blah.^{1}

and the second like

blah blah [1].

For AIAA publications, this is obnoxious because conference papers are like the first, but journal pubs are like the second. Instead of manually changing the position of the \cite{} command follow the convention in the .tex file

blah blah \cite{somedude2001}. (sentence+space+\cite{}+punctuation)

then invoke the natbib and natmove command like the first two lines suggest

\usepackage[numbers,sort&compress]{natbib}
\usepackage{natmove} % This command moves superscript references after the punctuation mark regardless of where the \cite command appears
\usepackage{hypernat}
\bibpunct{}{}{,}{s}{}{} % superscript citation

This functionality can also be achieved with the \usepackage[super]{cite} command but this overwrites the sort&compress and hypernat commands.

\bibpunct specifies all the delimiters and styles to be used with the reference numbers. Change the 's' to 'n' for inline numbers. Google it!

Aligning Axis Labels in OriginPro

2010-06-09T11:21:00.001-07:00

If you have multiple figures and you want all the axis labels to be in the same spot, copy the label from a finalized figure and paste it anywhere into the new one. Then hit ctrl+z and it will snap to the same location on the new graph.

OriginPro: How to label axis ticks in terms of pi

2010-06-09T10:41:00.000-07:00

If you're plotting numeric values from 0-2pi for instance and we want the x-axis labels to be in terms of pi (0, pi/2, pi, 3pi/2, 2pi) then open up the book containing your dataset and add a column.

Then enter the values of 0, p/2, p, 3p/2, 2p in the rows of the new column.

Then go to the plot and double click on the x-axis then the Tick Labels tab.

Change the font to Symbol and the "type" to Tick-indexed datas.

In the Dataset box select the column containing the axis labels and click apply.

Big fractions in Latex Array Environment

2010-06-08T14:20:00.000-07:00

Matrices and Arrays shrink fractions by default but sometimes we want to keep them large for display purposes. To force big fractions

replace \frac{}{} with \dfrac{}{}

this will make fractions their normal size in the display environment

Mathtype in Word 2010

2010-06-04T08:17:00.000-07:00

I'm starting to get acquainted with Office 2010 and like it so far but just as with the upgrade from 2003 to 2007, Mathtype doesn't work again! It doesn't show up in the ribbon and doesn't have permission to run macros.

Step 1: Integrating Mathtype with the ribbon

check out this link for a nice starting spot. It talks about integrating with Powerpoint and Word, but inevitably did not work with my install. He says to integrate with Word

Copy

\MathPage\MathPage.wll
\MathPage\MathType Commands 6 For Word.dotm

\Office14\STARTUP\

except my install directory didn't contain the .dotm file!
Instead download this file and save it to the STARTUP folder.

Step 2:  Enabling Macros so Mathtype actually works

Now the ribbon will contain Mathtype too but it will be disabled by default.  So I followed this link to enable macros.
It's a little different in 2010

File>Options>Trust Center>Trust Center Settings

I selected "Disable all macros except digitally signed macros."  Then restart Word and click the Enable Macros button in the
warning box

I'm sure Mathtype will fix this in later versions...and I'm sure they'll make us pay for it too.  For now, this works fine!

Manually Adding Packages to MikTeX

2010-03-01T13:06:00.001-08:00

A lot of journals supply class files. Many require adding additional packages to your installation. Inexplicably, they don't give much help installing. WTF? get with the program! Latex is supposed to be about the content, not the format. All the extra garbage ya'll require to install is the exact antipode of that idealism!

Anyway, this example will follow a revtex install over MikTeX in Windows
The download file will contain folders labeled bibtex, doc, source, and tex. Copy and paste these into the MikTex install folder (Program Files/MikTex X.X) and merge with existing. For MikTex to recognize the new packages go to
Start-All Programs-MikTeX X.X-Settings
(or Click Start → Programs → MiKTeX 2.8 → Maintenance → Settings)

Click on the button "Refresh FNDB"

I referred here to figure this out.

done!

Changing the Size of Delimiters in Multline Environment

2010-01-13T13:57:00.000-08:00

This is my first real problem with LaTeX. For long equations the multline enviroment works well to break them over two or more lines.

Lets say I want y=2+\frac{1}{1+\frac{1}{x}} displayed over two lines like

y=2
+1/(1+1/x)

then i would use (this example is too short an equation so LaTeX will spread it out too far, but the concept still applies)
\begin{multline}
2 \\ +\frac{1}{1+\frac{1}{x}}
\end{multline}

But lets say I want y=(2+\frac{1}{1+\frac{1}{x}})^{1/2} displayed over two lines like
y=(2
+1/(1+1/x))^{1/2}

I can use the \left( 'my equation' \right. and \left. 'my equation' \right) to close parentheses and manually break the equation like

\begin{multline}
\left( 2 \right.
\\
+\frac{1}{1+\frac{1}{2}} \left. \right)^{1/2}
\end{multline}

But if the height of the equation is the same for each line then this command will not match the size of the parenthesis.

to fix this problem use the \vphantom{} command. This command calculates the correct size for anything in the braces and hides the output - therefore the dual layer fraction will be used to match the size of the parentheses without displaying the extra terms. The easiest way is to just copy the offending terms into the braces and be done with it like

\begin{multline}
\left( \vphantom{+\frac{1}{1+\frac{1}{2}}}
2 \right.
\\
\left. +\frac{1}{1+\frac{1}{2}} \right)^{1/2}
\end{multline}

its sloppy and brutish to do it this way, but it works. A big WTF goes out to the multline command for not taking care of this automatically!

All the credit for this info has to be given to dnemoc who posted a shorter response to this question here. Thanks, that problem was really pissing me off!

Origin Plot Settings for HQ Graphs

2010-01-04T12:01:00.000-08:00

If you've hit this blog before you may realize I think Origin is a TERRIBLE piece of software to use. I have resolved to become proficient with it since there is a license available to me and my boss uses it exclusively. All I have to say is thank god I'm using LaTeX for most of my pubs because as far as I can tell Origin still gives you the middle finger if you want predictable results in MS Word. Anyway...here are the settings and steps that seem to work best for me

Open Origin and click on Tools>Options

Once you have a graph started under Format>Page

*UPDATE* one more thing...it seems that if you make graphs on different computers one more setting should be changed to ensure they are standardized. I think the disparity might be with the printer settings on different computers. The layer location on the graph needs to be manually set. Use

This seems to center the figure vertically well.
This is for a B&W publication that will fit 2 graphs horizontally for standard margins. Leave the rest alone

Then since I use LaTeX, you must export them into .eps so
File>Export
be sure Margin Control is on Page. That's important for all the axis to line up.

*UPDATE* Because Origin's terrible programming, it is important to export ONE GRAPH AT A TIME! Doing multiple exports at once can shift the position of labels and axis of the output file so that subfigures won't align properly. That was an hour of my life I'll never get back...thanks Origin. Good one :(

*UPDATE x2* Ok, amending some of the settings above to maximize the figure area. Set the page to 3.13 x 2 and then the layer dimensions look like

then make the axis major ticks to be 3pts. This will make the axis labels fill the entire page when the scales are as small as -9999 (4 characters from the edge of the axis to the edge of the page) for both x and y.

Then in LaTeX the command will look something like

% PRESSURE FIGURES
%===============================================================================================
\begin{figure}
% Requires \usepackage{graphicx}
\centering % centers everything on the page
\subfigure[]
{
\includegraphics[width=3.13in, height=1.75in]{Final_Figures/Pressure_dpdr}% include figure
\label{Fig. SubSec: Pressure_dpdr}
}
\subfigure[]
{
\includegraphics[width=3.13in, height=1.75in]{Final_Figures/Pressure_P}% include figure
\label{Fig. SubSec: Pressure_P}
}
\caption{Variation with vortex Reynolds number of \ref{Fig. SubSec: Pressure_dpdr} the radial derivative of pressure and \ref{Fig. SubSec: Pressure_P}. $\kappa=0.103$ for all plots.} % text to be included under all figures
\label{Fig. SubSec: Pressure}
\end{figure}
%===============================================================================================

Convert bibtex entries to \bibitem in Latex

2009-12-05T11:02:00.000-08:00

Some journals such as the Royal Society refuse to process .bib files for latex submissions. Seriously guys...get in the game! Instead they require the use of a bibliography command like

\begin{thebibliography}{3}
\providecommand{\natexlab}[1]{#1}
\expandafter\ifx\csname urlstyle\endcsname\relax
\providecommand{\doi}[1]{doi:\discretionary{}{}{}#1}\else
\providecommand{\doi}{doi:\discretionary{}{}{}\begingroup
\urlstyle{rm}\Url}\fi

\bibitem[{Anderson(2001)}]{Anderson2001}
Anderson, J. 2001 History of high-speed flight and its technical development.
\newblock \emph{AIAA J.}, \textbf{39}(5), 761--771.

\bibitem[{Anderson(2003)}]{Anderson2003a}
Anderson, J. 2003 \emph{Modern compressible flow: with historical perspective}.
\newblock McGraw-Hill.

\bibitem[{Stodola(1905)}]{Stodola1905}
Stodola, A. 1905 \emph{Steam turbines}.
\newblock D. Van Nostrand.

\end{thebibliography}

where all the bibliography entries are listed as \bibitems. Fortunately I found a way to convert all your entries in the .bib file to \bibitems from here.

It says
1) Generate your bibliography with regular bibtex commands first (run latex, bibtex, latex)
2) Open up the new .bbl file
3) Ctrl-A to select all, Ctrl-C to copy
4) Paste over the \bibliography{} command
5) rerun Latex!

Bibtex makes all the \bibitems automatically, but behind the scenes.

Blue Border on Blackberry Apps

2009-10-27T10:52:00.000-07:00

I had a blue border and permanent keyboard locked in portrait view (wouldn't rotate) with the touch screen being slightly off (didn't accurately recognize where my finger was) after reinstalling Google Sync. To fix it I went to

options>applications>highlight Google Sync>click menu>click Disable Compatibility mode>pull the battery

This worked for me and with this app. I've seen the same blue border on a friends phone with YouMail and with the QuickPull Scheduler. I don't know if it would fix those problems as well, but it's definitely worth a shot! I gathered it was a Java problem but i dont know. It made the apps unusable.

Blackberry/Outlook/Gmail sync

2009-10-27T09:09:00.000-07:00

IF Google was willing to include contact sync with their desktop calender sync program life would be much easier. I don't know why mobile sync includes both but desktop doesn't. I also don't know why gmail contacts info fields aren't the same as outlook. I think outlook has a better system simply because the interface and extendability are a little better than gmail. Anyway, here's how I sync my Outlook to Blackberry Storm to Google Calender/Contacts to Outlook.

First thing Backup!

This method is free and works without making duplicates. It's not perfect and i'm only writing this because over the last 24 hours I've had to restore all my contacts and fix my calendar about 3 times because the new OS 5 came out and i couldn't help but screw around...and ruin everything.

1. Install Google Calendar sync on the PC (http://www.google.com/sync/pc.html)
- sign in and select 2-way sync

2. Blackberry Desktop Manager
-click on synchronization-configure synchronization (in the left pane)
-click on synchronization (in the right pane)
-check what you want but UNCHECK the sync Calendar box
This step make sure the contacts get synced to outlook. Because google sync for pc doesn't sync contacts there's no problem of duplicates.

This still means you have to connect to the host PC with either the cable or bluetooth...i dont have bluetooth so it's not entirely wireless.

3. Install Google sync mobile to the blackberry. (http://www.google.com/sync/blackberry.html)
This automatically puts contacts and calender events on Gmail and Google Calendar which is synced with Outlook with step 1.

I'm set up so the blackberry is the primary device. This is because i actually sync across multiple computers too and the blackberry can be the go between.

Verticle Scroll bar Missing in Word 2007

2009-05-27T10:26:00.000-07:00

Sometimes the vertical scroll bar in Word 2007 disappears for no apparent reason. To turn it back on go to

the Windows button in the top left corner, Word Options, Advanced, Display

then check 'show vertical scroll bar'

I don't know why this is even an option.

Remote Desktop with Mulitple Monitor Support

2009-05-15T08:23:00.000-07:00

This post is basically a bookmark for myself so I can quickly find the links I used to figure out how to remote desktop over multiple monitors.

for multiple monitors of the same size and resolution the 'run' command is

mstsc /span

this was found here

for monitors of different resolution

mstsc /W:2860 /H:1050

where 'W' is the total width of both monitors (2860=1680+1280) and 'H' is the height of the smallest monitor.

this was found here

The problem is that the computer i'm trying to log in to has two identical monitors, but the computer i'm logging in from does not. The /span command doesn't seem to work correctly.

Making bibliography styles for Latex

2009-05-04T11:19:00.000-07:00

This is actually much easier than I expected.

Open a command prompt and type
latex makebst

this opens a dialog where you fill out a bunch of formatting questions. It creates a .dbj file and ask if you want to run this 'batch job now.' Clicking yes generates the .bst file you'll need for the \bibliographystyle{} command

The output files are in C:\Users\Username

Grouping Nomenclature in Latex

2009-04-27T12:10:00.000-07:00

See the other post about nomenclature for general usage

the \nomenclature command has some other options given by

\nomenclature[]{}{}

the is replaced by a letter or number to define groups in the nomenclature. You can create group titles by following the 'subgroups' section. If you JUST want to group things together, just use numbers like

\nomenclature[1]{$A$}{The nozzle cross-sectional area}

\nomenclature[1]{$A^*$}{The throat area}

\nomenclature[2]{$\varepsilon$}{The area ratio, $\varepsilon=A^*/A$}

\nomenclature[1]{$M$}{The Mach number}

\nomenclature[2]{$\gamma$}{The ratio of specific heats}

to group the roman and greek letters seperately.

For the aiaa-tc class from aiaa.org i had to do this because they had subgroups defined in the class. There subgroups are

b = subscript
c = convention
g = Symbols
t = superscript

This is usually dependent on \usepackage[]{nomencl} or \usepackage[noprefix]{nomencl} and i haven't messed around too much with other classes, but the aiaa-tc class overwrote the option anyway so i had to use numbers instead of letters for everything.

Nomenclature in Latex Made Easy

2009-04-27T07:44:00.000-07:00

As pretty as Latex documents are, for 40 years of development you'd think they'd have made all this automatic. I'm using Texmaker to generate the Latex output and a BATCH file to generate the nomenclature. The Latex syntax you'll need can be found here.

In the directory where the Latex file is located create a BATCH file (a text document with the extension changed to .bat) with the following command

makeindex.exe filename.nlo -s nomencl.ist -o filename.nls

replace 'filename' with the name of the tex file. For example if the tex file is called 'glossary.tex' then the BATCH file command would look like

makeindex.exe glossary.nlo -s nomencl.ist -o glossary.nls

The process is 3 steps,
1) Run Latex on the document to identify the nomenclature,
2) build the nomenclature by double clicking on the BATCH file,
3) print the nomenclature to the output file by running Latex on the document again

The process should similar for generating indexes and glossaries except the Latex syntax is a little different. They all use the makeindex.exe command so they all need different identifiers.

Figures, Subfigures, and Table Syntax in Latex

2009-04-24T09:26:00.001-07:00

UPDATED*  Now I use the package \usepackage{float} for figures and tables.  It allows easy manual editing of how the figures are labeled and works for both tables and figures.

The code looks like

\usepackage{float}                      % figure floats

\usepackage{subfig}  % package for subfigure formatting \renewcommand{\thesubfigure}{\alph{subfigure}}% removed \textbf lest Fig. 1a is bolded in the text \renewcommand{\thesubtable}{\arabic{subtable}}% removed \textbf lest Table 1 is bolded in the text

%===============================================================================================

% GRID REFINEMENT %===============================================================================================       \begin{figure}      % Requires \usepackage{graphicx}      \centering % centers everything on the page     \subfloat[\textbf{a)} Axisymmetric perturbations ($q=0$)]     {         \includegraphics[width=3.13in, height=1.75in]{figures/Ch_LNP_of_BV/LNP_Axisymmetric_Grid_Refinement}% include figure     \label{Ch:LNP of BV Fig:Axisymmetric Grid Refinement}}     \subfloat[\textbf{b)} Asymmetric perturbations ($q=1$)]      {         \includegraphics[width=3.13in, height=1.75in]{figures/Ch_LNP_of_BV/LNP_Asymmetric_Grid_Refinement}% include figure     \label{Ch:LNP of BV Fig:Asymetric Grid Refinement}}         \caption{Grid refinement for $\alpha=3$, $z=1.5$, $Re=10,000$, and $\kappa=0.1$} % text to be   included under all figures         \label{Ch:LNP of BV Fig:Grid Refinement}     \end{figure} %===============================================================================================

OLD WAY USING SUBFIGURE


\documentclass{article}
%   PREAMBLE
%=================================================================================================
\usepackage{amsmath}    % math format package
\usepackage{booktabs}   % professional tables package
\usepackage[final]{graphicx}   % figures package
\usepackage{subfigure}  % package for subfigure formatting
%=================================================================================================

% you must include \usepackage{graphicx}.  this has 2 options - [final] for final drafts - [draft]
only gives a place holder and doesn't render the image.

% for pdf documents the image must be in pdf form. for dvi the image must be in eps.

% to use figures in a separate subdirectory use "folder_name/file_name"

% it's not perfect because it is touchy about the files it can use.  you could use a dummy eps
file so the dvi compiler will work. or you do the figures only in pdf and don't include them until
done.

% \includegraphics[width=?]{} gives options to change the properties such as width, leaving it
blank uses the original sizes

% some of the journal styles cause problems with this for some reason
%=========================
% THIS IS A SINGLE FIGURE
%=========================
\begin{figure}
% Requires \usepackage{graphicx}
\centering % centers everything on the page
\includegraphics{Final_Figures/S=50}\\ % include figure
\caption{$\lambda=50$} % text to be included with figure
\label{fig:lambda=50}
\end{figure}

%=================================
% THIS IS A FIGURE WITH SUBFIGURES
%=================================

\begin{figure}
% Requires \usepackage{graphicx}
\centering % centers everything on the page
\subfigure[] % text to be included with figure 1 goes inside '[]'
{
 \includegraphics[width=2.5in, height=1.75in]{Final_Figures/S=50}% include figure
 \label{fig:1lambda=50}
}
\subfigure[] % text to be included with figure 2 goes inside '[]'
{
 \includegraphics[width=2.25in, height=1.75in]{Final_Figures/S=100}% include figure
 \label{fig:1lambda=100}
}
\subfigure[] % text to be included with figure 3 goes inside '[]'
{
 \includegraphics[width=2.5in, height=1.75in]{Final_Figures/S=200}% include figure
 \label{fig:1lambda=200}
}
\subfigure[] % text to be included with figure 4 goes inside '[]'
{
 \includegraphics[width=2.25in, height=1.75in]{Final_Figures/S=400}% include figure
 \label{fig:1lambda=400}
}
\caption{Analytical and numerical solutions of the eigenfunction equation with
$a_0 = -\cos \left(\frac{1}{2}\pi y\right)$, $\xi=5$, $n=1$, $\omega_n t=\pi/2$, $x/l=0.5$
and a Strouhal number of (a) $\lambda=50$  (b) 100   (c) 200   and (d)
400. The agreement between numerics and asymptotics continues to improve with
successive increases in the Strouhal number despite the highly oscillatory nature of
the solution and the radical reduction in spatial wavelength.} % text to be included
under all figures
\label{fig:Solution with several Lambdas}
\end{figure}

% =====================
% THIS IS FOR TABLES
% =====================
% be sure to use \usepackage{booktabs} in the preamble

% tables and tabular are different things.  Tabular gives you the format of a table,
but the
\table command is required to actually identify it as a table in the document with automatic
cross-referencing and whatnot.

% the tabular command has the format \begin{tabular}{ x y z }.  {x y z} are replaced
by the cell formatting 'l' for left 'c' for center 'r' for right.  it can appear as
{|x|y|z|} to give vertical lines

% to span multiple cells use
\multicolumn{number of cells to span}{cell formatting}{What appears in the cell goes here}\\

% the '\\' command moves to the next line

     \begin{table} % begin the table identification command
 \addtolength{\tabcolsep}{-4pt}
     \centering % ensure it centers on the page.  modify this if centering is not wanted
     \caption{Comparison between numerical and asymptotic solutions for both Type I and Type II.
Here $\lambda=10$, $\varepsilon=10^{-3}$, $n=0$, $a_0=-\cos(\frac{1}{2}\pi x)$, and $b_0=(2n+2)F'$}
% this is the text appearing with the table - description
     \label{table:Comparing the Types} % this is the label for cross-referencing

         \begin{tabular}{c|c c c|ccc|c} % this begins the table format
             \toprule % bold horizontal line command
& \multicolumn{3}{c|}{Type I}& \multicolumn{3}{c|}{Type II}& Numerical\\
\cmidrule(){2-8}
$x$ &   $f^{(2)}$ & $f^{(3)}$ & $f^{(4)}$ &
$f^{(2)}$ & $f^{(3)}$ & $f^{(4)}$ & $f^N$\\
             \midrule % horizontal line command

      0 &          1 &          1 &          1 &          1 &          1 &          1 &          1 \\

   0.05 &  0.8676644 &  0.8674551 &   0.867459 &  0.8676685 &   0.867459 &   0.867459 &   0.867459 \\

    0.1 &  0.5188763 &  0.5186301 &  0.5186429 &  0.5188833 &  0.5186428 &  0.5186429 &  0.5186429 \\

    0.2 & -0.3939219 & -0.3935576 & -0.3935347 & -0.3939428 &  -0.393534 & -0.3935347 & -0.3935347 \\

    0.3 & -0.7672007 & -0.7662141 &  -0.766217 & -0.7672949 & -0.7662147 &  -0.766217 & -0.7662171 \\

    0.4 & -0.2589166 & -0.2585327 & -0.2585659 & -0.2589812 & -0.2585645 & -0.2585659 & -0.2585659 \\

    0.5 &  0.3586916 &  0.3582113 &  0.3582018 &  0.3589124 &  0.3581976 &  0.3582018 &  0.3582018 \\

    0.6 &  0.2062112 &  0.2061807 &  0.2061778 &   0.206496 &  0.2061744 &  0.2061778 &  0.2061778 \\

    0.7 & -0.1574955 &  -0.158249 & -0.1582805 &  -0.158256 & -0.1582757 & -0.1582805 & -0.1582805 \\

    0.8 &  0.0345428 &  0.0357267 &  0.0360158 &  0.0356684 &  0.0360156 &  0.0360158 &  0.0360158 \\

    0.9 & -0.0049035 & -0.0104253 & -0.0091613 &  -0.008804 & -0.0091688 & -0.0091612 & -0.0091613 \\

   0.95 & -0.0000306 &   34513650 &  0.0008021 &  0.0006995 &  0.0008034 &  0.0008017 &  0.0008015 \\

   1.0  &          0 &  $\infty$  &          0 & -0.0000111 & -0.0000168 &  -0.000017 & -0.0000339 \\

             \bottomrule % horizontal line command
         \end{tabular}

 \end{table}

\end{document}

Be sure to put the \label{} command below the \caption{} command for figures and tables. See this post for more on this point.

I used this link to learn about subfigures and this and this for tables. Also, i found a great editor that exports excel data to latex format here.

This is where i figured out how to put verbatim code in Blogger

Path Errors in Texmaker on 64 bit

2009-04-23T06:57:00.000-07:00

Hopefully the Texmaker developers will address this problem. Miktex installs in the Program files (x86) directory on a 64 bit machine. Texmaker defaults to the regular old Program files directory when looking for the executable dvi, pdf, ps viewers. You have to manually set up the correct paths.

Texmaker is pretty good so far...i would like to try Kile, but i refuse to install Linux just to test it out. If Texmaker impliments a couple more things, they would be formitable. Right now, no tex interface is perfect. Some one smarter than me should take the one or two outstanding features in each and develop a new one.

Reverse DVI Search in TeXmaker

2009-04-22T15:21:00.000-07:00

Since my Winedt trial expired, I switched to TeXmaker. It's got a pretty interface, but needs some tweaking to make work right...for instance installing a dictionary...

I had a hell of a time figuring out how to make the Reverse DVI search work in TeXmaker. Here is the solution

in Texmaker Options>Configure Texmaker, replace the LaTex commands with
latex --src -interaction=nonstopmode %.tex %0 "%N%T"

This tells Texmaker to generate "source links" so Yap can link backwards.

in Yap click View>Options>Inverse DVI search>New
Give it a name, navigate to the Texmaker.exe file and add
"%f" "%l"
be sure to INCLUDE the quotation marks or else it won't work (even though it lets you click 'ok' without them)

Set Texmaker as the default and you're done.

*UPDATE See the discussion in the comments section for some good suggestions

Exporting References from Endnote to BibTeX

2009-03-21T11:07:00.000-07:00

Endnote has a bibtex export style. I'm not sure why they bothered. It doesn't work, it's never worked and from the looks of things this and a couple other issues with Endnote are never going to work. The articles and third party software i've tried just don't work or are a real pain to impliment.

I use JabRef for my bibtex reference manager. They seem to be on top of it. Besides the ease of use and functionality of the software in the first place, they seem to have an easy and built-in (sort of) way to export correctly from Endnote.

Step 1
In JabRef click Tools-unpack Endnote Filter Set (this is a zip file so put it anywhere)

Step 2
Open the zip and read the "read me" file. It has instructions to go from Endnote to JabRef and JabRef to Endnote. For our purposes copy BibTeX Export to JabRef.ens and BibTeX Export to JabRef2.ens into the styles folder
(something like C:\Program Files (x86)\EndNote X2\Styles).

Step 3
In Endnote click on file-export. Then in Output Style, select Use Another Style. In the search box type jab and it'll bring up the two new styles. You can see the difference in the preview pane, but JabRef.ens is the one we want rather than JabRef2.ens.

Step 4
Save as a text file

Step 5
rename the file from .txt to .bib and open it in JabRef. You can also import. WARNING!! this will FAIL if any of the references has a bibtex key already that contains a space (or punctuation maybe too). Any key like this needs to be deleted or else it won't work.

Step 6
In JabRef, select all and click the magic wand to generate new keys for all entries. DONE

I actually exported some references from JabRef an copied all my Endnote exports into that file. There's some stuff in the JabRef export that is commented out. I don't think it needs them, but just incase.

Properly labeling Tables and Figures in Latex

2009-03-20T14:51:00.000-07:00

As with all things in Latex; it can be done, but it might not make sense.

If you have tables and figures throughout your document you might come across the problem that your \ref{} command considers tables and figures the same thing and therefore gives the wrong number. For instance if your document has tables and figures appearing in the following order

Table 1
Table 2
Figure 1

and you try to reference Figure 1 you might actually get a "3" instead of "1." This happens if you label your figures and tables wrong. To avoid this always make sure the \label{} is AFTER the \caption{} for both tables and figures. That should make everything work fine.

Remember

\caption{}
\label{}

Error Plots in Graphpad

2009-03-11T16:59:00.000-07:00

Graphpad is designed for statistical analysis. It has some really complicated analysis tools for those apps. However it is NOT easy to wade through all that mess to do something simple like take one column and subtract it from the other.

To do simple stuff like that click
>Analyze
then select the data sets to be included then click
>Remove baseline and column math (Transform, Normalize list)
>Selected column(s)
Click the column you want as the base (for instance the column containing the exact solution)
>Calculation (pick your operation)
>Create new graph of the result

that's it. I'm not sure how it handles negative numbers exactly. you might have to create another column or do it by hand.

Alignment for Labels in Graphpad

2009-03-11T15:31:00.000-07:00

I don't use the built in axis or title labels. Instead I use Mathtype. Most of my labels require numbers and symbols anyway so i stay consistent by using it throughout.

The picture below shows where i like to position the Mathtype labels. The black squares always align to an edge.

Sizing Graphpad Plots

2009-03-11T14:38:00.000-07:00

In order to generate plots that are always the same regardless if they appear in print alone or with other plots in the same figure you should build a layout.

My first post on Graphpad has my Preferences so the numbers here are based on those settings, but it might not matter.

First select the type of plot, import the data, and let it autogenerate the plot. Then go in and fine tune the details (labels, line type, legend position...). Start a new data table and import the next set...so on and so forth.

First thing to know is that for multiple graph layouts each graph appears as 86% of the original size. Therefore for single graph layouts (yes create a layout for a single graph too because the sizing options are different) need to be resized to 86%. If you don't use my preferences make a horizontally spaced layout and put two graphs in it. Then double click on one to see the size in percent. By default it autoadjusts and tells the percentage. Different journals have different margins so you'll be figuring it out on your own more often than not.

This seems to work regardless of how you make and position axis labels, titles, legends. If they don't line up, there is an "align axis" option and it'll work fine. Also, don't forget the option Right Click equalize graph size.

IF you're using Word for the final document...there is a button to send directly to Word or Powerpoint. This works great(ish). For layouts with several figures, this couldn't be easier. 1 button click. For single graphs, Word enlarges them slightly so you have to fumble with the sizes to make all your plots consistent.

For LaTeX users, I would still bother with multigraph layouts. The subfigure package for LaTeX will do it for you, but if you don't get the graphs exactly the same size, it'll be finicky. The graphpad layouts are easier.

Either LaTeX or Word users, i'd suggest exporting the layouts to something else rather than using the auto send button for Word. Clearly, LaTeX users have to do this. Graphpad makes this easy by giving the option to just export only the graph and not the whole white page for pdf and eps. There are other options for other image types. I export to pdf and include \usepackage{subfigure} for LaTeX. For Word I build Layouts and export as either EMF or TIF. Resolutions can be changed if need be.

UPDATE:
If you want to use the subfigure package in LaTeX, then I recommend exporting each figure individually and using the following settings:

Double click on the graph (not the layout graph)
Size: Wide
Frame: 3.13in wide X 2.09in high

click export, select format, and set
Size: Make width 4.24

That gets the whole figure neatly framed

in LaTeX
\includegraphics[width=3.13in, height=2.09in]{"figure file"}

That gives a nice side by side

Graphpad vs. Origin for Scientific Plotting

2009-03-11T14:23:00.000-07:00

I had been using Origin to generate plots mostly because that's what my adviser used. I absolutely despise Origin. I complain that I expect things to "just work." Origin caused me so many headaches during my Master's thesis that I was desperate to find another option.

I found Graphpad. I've been messing around with the trial (nothing good is ever free) and have decided it is superior for a few simple reasons. First it integrates Mathtype so i can guarantee the size of my fonts are always consistent. In Origin you had to copy in Mathtype as images and resize inside Origin, but the point size differed if it had subscripts or superscripts. There is an "align" option in Graphpad that guarentees alignment of axis in the final result. This is also useful for inset plots or denoting graph a, b, c,...etc. and making sure that everything lines up correctly.

Also origin would leave large white areas around the outside of the plot unless you changed the preferences. Importing a plot made with someone elses settings would almost always screw things up. Importing to Word was also a nightmare with crashes, but mostly with alignment problems. Graphpad has a layout feature that builds the final plot for you. For instance if I want two plots side by side, I just go to that layout and drag and drop the previously (automatically) generated plots into the placeholders.

If you have multiple plots that you want to ensure are identical in format, there's a magic want tool for that. It's not more than 3 clicks. If you are importing data from multiple files or only certain columns or rows, the filter tab has everything you need. You can view EVERYTHING before it is generated - the data before it's imported or the graphs before they are formatted.

Origin is more powerful. Graphpad appears to be made for business rather than science. It can't do 3D plots or some of the other fancy things Origin can do, but for simplicity, predictability, intuitive design, and repeatability Graphpad is superior by far. Consider that I went from downloading the trial to figuring out how to set up preferences to importing data to generating professional high quality graphs in about an hour with Graphpad. I've been using Origin for over 4 years and still haven't figured out ways around the problems I've faced.

These are the Preferences i use for scientific plots

AXES
height -1.75
shape -wide
frame -no frame
thickness -1pt
ticks -outside

Default color scheme black and white

FONTS
main -regular times 14
axis titles -regular times 12
numbering -regular times 12
legend -regular times 12

SYMBOLS AND LINES
line thickness -1/2pt

FRAME
3.13w
1.75h

Network Mapping with XP and Vista

2008-09-21T07:04:00.000-07:00

Getting Vista and XP to play nicely on a network is certainly not always a simple task. My situation was: An XP machine with a shared folder I wanted map to a Vista machine on the same network, but a different subnet. I 'think' the following instructions are just a generally good idea if you are looking to communicate between XP and Vista regardless of the subnet. It's one of the first things I do when I reformat.

You have to change Vista's security policy to recognize and allow communications with XP's infrastructure. The local security policies can be found in Control Panel>System and Maintenance> Administrative tools.
Open Local Security Policies and expand the Local Policies folder on the left. Then click on Security Options.

Toward the bottom, double click on Network Security: LAN manager Authentication Level and set it to Send LM & NTLM responses.
It doesn't tell you to restart, but I always have to before it actually takes the settings.

I think you would probably need to do this on the Vista computer if it were hosting the shared folder. FYI, I spend probably 8 hours trying to figure this out...too much time for a 30 second solution!

Also, if it still isn't working try shutting off Windows firewall and testing it. The firewall seems to get in the way when you're on different subnets - even if it says file sharing is allowed through the firewall. If your network has an external firewall it might be OK to leave it turned off...then again it won't be my problem if something goes wrong because the firewall was shut off!

Removing password request after returning from standby

2008-09-16T18:24:00.000-07:00

Here's a link explaining how to remove the password request in Vista when returning from standby or hibernate.

http://www.mydigitallife.info/2008/06/29/disable-windows-xp-and-vista-from-asking-for-password-to-unlock-on-wake-up-resume-from-sleepstandby/

That's the long way around it. I found that you can just open the power settings and there's a link to the right spot. See the picture
then click on the 'do not require a password' radio button

Mathtype Quick Keys

2008-09-14T19:56:00.000-07:00

There's a bunch of quick keys in Mathtype for actually entering equations. If you find the symbol you want to enter from the menu (shortcuts don't work) and hold the mouse over it the quick commands will show up at the bottom of the window. These quick keys from a word document that actually insert an equation.

insert an inline equation: alt+ctrl+q
insert a numbered equation: alt+shift+q
insert a display equation: alt+q

Mathtype error in inUILibCls.Error

2008-09-14T19:15:00.000-07:00

I got this error after using Mathtype integrated with Word 2007 for a long time and I couldn't figure out why. This error is because Mathtype needs a defined Section to begin numbering. If there is nothing but text above the location of the first numbered equation it seems to assume it is in Section 1 and begins numbering from 1 without a problem. This error occurs when you have quickparts above the first numbered equation. I got this error when I inserted a cover page from the insert menu. The built in cover pages are made entirely of quick parts. To fix this, just manually define a Mathtype Section or Chapter break (see picture) before the first numbered equation. For a cover page the break actually sits above the abstract...for some reason that's what Word thinks is the last line of the page.
I usually put them on the first line of text - after the headings. You can see the sections breaks in the text if that button that shows all the formatting is pushed (I don't remeber what its called, but it looks like a carriage return symbol), but they don't print.
If you have them on the same line as a heading, they will show up on the auto generated table of contents and you'll have to delete them out.

Quick Keys for Making a 'New Folder' in Windows Explorer

2008-09-04T09:38:00.000-07:00

To make a 'New Folder' in Windows Explorer the key commands are

alt+F,W,F (capitalization doesn't matter)

That is the quick command for File>New>Folder so it's not really a quick key, but it's faster than right-click..wait..new..wait..Folder.

It doesn't work on the desktop unless Windows Explorer is open to the Desktop. You could bind a special key to make it just one button if you were so inclined.

Equations in Blogger (and other HTML) Made Easy

2008-08-29T12:03:00.001-07:00

With a little help from this blog, I finally figured out how to create aesthetically pleasing equations in blogger or any HTML site without a big hassle. His site was a big help, but I still had to fumble around to get it working. Hopefully this post will clear up any questions.

The problem is that Firefox renders MathML in .xhtml, but Blogger publishes in .html so there's a compatibility issue. The solution is to input a Java script in the html that handles the equations for you. The script is thanks to this site. It is licensed under the GNU. You will need Firefox or IE with mathplayer.

Step 1: Get the script

Go to http://www1.chapman.edu/~jipsen/mathml/asciimathdownload/ for the original host OR go to http://students.utsi.edu/jbatters/scripts/asciimathml.js for one that renders the equations in black instead. If the second link opens the text version, just save the page as .js.

Step 2: Host the script

The script has to be hosted somewhere for it to work. I hosted it on my personal website (the link above) but it can be hosted anywhere. Google Sites should work. OR I don't mind if you just use my script that's already hosted.

Step 3: Call the script from Blogger. Add the html/java script box to your layout. Follow the arrows

Then add the code to your html/javascript box. Click the 'edit' button shown from the Layout edit screen. Paste the following in the box.

That script points to my hosted file. If your hosting your own replace 'students.utsi.edu/jbatters/scripts' with your host location

Be sure to SAVE.

That's all the setup. Now it will work for ANY new post on that blog. For your own personal website you can point to the java script hosted elsewhere with the same command by pasting the script text between the 'head' tags at the top of your html file or you can use the following command with the java script saved in the same folder as your html file.

<head>

<script type="text/javascript" src="ASCIIMathML.js"></script>

<title>your title...</title>

</head>

Step 4: Entering equations in Blogger

See the example for a simple equation

The script recognizes the text between the accents (the button above the 'tab') as math. You can get some syntax from here. It recognizes many LaTeX commands. Example

`int_(0)^(1)(x^2+2x)dx`

Since this works with MathML it might work directly with Mathematica, Mathtype, etc. I'll write up instructions to export TeX from Mathtype. That should work if the MathML thing doesn't.