Quantum Forest

Spotty posting is a clear symptom of either being sick with internet or just too busy in real life. In my case is mostly the second; this is the time when I have overlapping teaching of regression modelling and introduction to tree breeding. The other thing I have been doing is completing project reports (two gone!) and playing with some data for a journal manuscript. Overall, I have written more this month than at any other time that I can remember.

On the contrary, I have been posting very little to this site and a few posts to http://trendecarga.com. Stating the obvious, the more I write offline the less I write online; and I have been feeling the urge to complete a series of pending writing projects.

A few weeks ago I submitted a paper that has been in that limbo-like close-to-finish for three years. As soon as I finish teaching in three weeks time, I will complete a second paper and start writing a third one. This will take a toll on this site, but one does not get brownie points for blogging or playing with HTML.

There will certainly be some updates to this site (although most likely not in the blog part) as well as a new design for http://plustree.com. On the latter I have been slowly working in a new template and CSS file, aiming for a cleaner and simpler look.

Re-stating the obvious, real life has much wider significance, it is richer and more meaningful than any web site. Do not let anybody convince you of the opposite.

One of the issues of living immersed in another language and culture is the slow — or not so slow — deterioration of the command of the original language. There is not enough practice, not enough interaction to keep it healthy and alive. To avoid this, I will be writing some posts in Spanish (hopefully) with a friend of mine.

Because this is an experiment, and quite different from the contents of this site it will be under a different name: Tren de Carga (Freight train). The name originates from a verse by Chilean poet David Turkeltaub, which says ‘vienes tarde otra vez, como tren de carga’ (you are late again, like a freight train). I am not sure why, but this verse has been stuck in my head for many years and this is the first time I have found a use for it.

Eleven of September again, but I am not thinking of 2001—which was terrible, I agree—but of 1973. I always stop for a minute to remember as an old political campaign said: without hate, without fear, without violence. I was six and still remember.

I am preparing lectures, really busy. I took the students after more than a semester and most of them appear to be clueless. I am going over a general review, again, because they have to learn. Always listening to energetic music when preparing lectures or running analyses. This time is Rammstein’s ‘Sehnsucht’. But, anyway, I also like to go back a few years and Yes’s ‘Close to the Edge’ (yes, from 1972) is now coming from the headphones.

I start writing most of the ideas in Writeroom and then copy the lot to TeXShop. Packages included in the document preamble for the notes:

%! program = pdflatex
\documentclass[11pt]{article}
\usepackage{pslatex}
\usepackage{amsbsy}
\usepackage{graphicx}
\usepackage{geometry}
\geometry{a4paper}

Although the Statistics course uses Mendenhall & Sincich’s ‘Regression Analysis’ as the basic text, I am using quite a few other references:

• Quinn and Keough’s ‘Experimental design and data analysis for biologists’.
• Searle’s ‘Linear models’.
• Harrell’s ‘Regression modeling strategies’.
• Steel and Torrie’s ‘Principles and procedures of statistics’ (the first edition!).
• Hamilton’s ‘Regression with graphics’.
• Neter and Wasserman’s ‘Applied linear statistical models’.

In addition I am also preparing grant applications, so if you try to contact me please be understanding: I will not read my email (or act upon any emails) until Wednesday next week (around September 20^th).

Moved to King Crimson’s ‘Discipline’. Now listening ‘Elephant talk’…Talk is ony talk…

Following my Another mac update post I will cover a few small applications that have made a big difference on the way I do things.

DigitalColor Meter: One of the common tasks that I perform when updating my web sites is to change the CSS files. I keep the same templates, but I do play a fair amount with the design; particularly colours. When I was using mostly a PC I would find a nice colour in a page, take a screenshot (with Shift-Print Screen), open the screenshot in The Gimp and use the color picker to get the hexadecimal representation of the colour. Moving to the mac I had the temptation of doing the same, but using Grab (which comes by default under Applications, Utilities, Grab.app) and The Gimp for mac or—a simplified version of it—Seashore. Then I discovered that the mac ships by default with Digital ColorMeter (under Applications, Utilities, DigitalColor Meter.app) a small piece of software that lets you hover above any screen and get the hexadecimal color. One can even modify the aperture size, to change the number of pixels being included in the average colour measurement. Now it is simply ‘Hover with my mouse, Shift-Command-C’ and I get the hexadecimal values.

Dictionary: The other one is simply Dictionary. Yes, I have a problem, I like dictionaries and related books. I do own a copy of ‘Webster’s New World’, ‘Oxford Advanced Learners’, ‘Roget’s International Thesaurus’, ‘Brewer’s Dictionary of Phrase & Fable’ and ‘Fowler’s Dictionary of Modern English Usage’. This is covering English only. I do have a similar set of books in Spanish, starting with ‘Diccionario del Español Actual’ by Manuel Seco, Olimpia Andrés and Gabino Ramos (4,666 pages), followed by a large number of smaller dictionaries and books on the use of Spanish. I used to go and look for words all the time, but now there is Command-Spacebar Di (in Quicksilver) and the word. That’s it and I love it. I now only need something like that for Spanish.

Still trying on writing tools

Writeroom keeps growing on me, but I need to store all the bits and pieces that I am writing for different reasons. Although I tried Mori (from the same makers as Writeroom), I still do not feel quite comfortable with it. Journler, on the other hand, it is a nice journaling application, although its full screen mode is still a bit clunky.

So, I wrote my first ever piece of Applescript:


tell application "WriteRoom"
set docName to the name of front document
set docContent to the text of front document
end tell
tell application "Journler"
set newEntry to (make new entry with properties {name:docName, plain text:docContent})
end tell


Writeroom assigns automatically the content of the first line as document name. Thus, the script grabs the name and content of a WriteRoom window and creates a new entry in Journler with that title and content.

Yes, I did buy a copy of Copywrite before, but I can not make it fit in my workflow yet. I will keep trying it for a while, and see if it ‘just clicks’ with me in the near future.

After a false start, I am again putting some of my bookmarks in del.icio.us. I will probably add the tags (newish term for old-fashion keywords) to blog posts too.

Playing with cream

Paul Ford’s comments on Amish computing certainly hit a soft spot on me. I do miss Wordperfect 5.1! It was back to simpler times when using computers was certainly much more productive for me. Multitasking is a nice feature to have when strictly necessary, but not all the time.

Next year I need to spend a fair amount of time writing lectures and I am certainly tempted to ‘going back to basics’. Most of the text that I need to prepare is not highly complex, so I am thinking of writing at least the first drafts in text files with a simple markup. The most humanly readable markup is probably Markdown. Once the text is in Markdown it can be easily converted into html (e.g. using the Markdown dingus, and adding the ‘html’ and ‘body’ tags to get a complete page) and from there to other formats like LaTeX or MSWord. If I decide to go for a longer document probably LaTeX would be the way to go.

I have been playing with Cream, the VIM mode for dumb users like me (another distraction). I hope to slowly learn a few tricks at a time to become a more proficient VIM user, but that is not a real priority. It is a really nice editor mode!

I installed the vim-latex suite, which seems to add pretty good latex support to VIM/Cream, but it seems to override some of the Cream configurations (e.g., F9 is not code folfing/unfolding anymore). It seems to be a matter of getting used to that though. Anyway, I will not need it for the first version of the documents.

Making more changes to Tim’s site

We have had a few problems to have the PDF file of Tim’s book indexed by search engines. My theory is that engines aren’t very happy with Textpattern’s internal links (of type http://mysite.com/file_downloads/2) for a PDF file. Today we changed it to something more explicit like http://mysite.com/bookfiles/file.pdf. Actually, the story was not as simple as that. When first trying to use the new code we ran into a ‘missing page’ problem, which I traced back to a problem with the .htaccess file. I dropped a <IfModule mod_rewrite.c> condition from the file, leaving it like below and it just works.


RewriteEngine On
RewriteCond %{REQUEST_FILENAME} -f [OR]
RewriteCond %{REQUEST_FILENAME} -d
RewriteRule ^(.+) - [PT,L]
RewriteRule ^(.*) index.php


Next year I have to prepare thirty two lectures; eight for an introductory class and twenty four for a 400 level elective class. The first will be the only exposure to tree breeding for many people, so I need to prepare a sort of ‘tree breeding survival kit’. The elective class will be taken by Bachellor and Masters (with extra course work) students, so there is much more room to explore topics.

One of the issues when preparing teaching material (and presentations) is what is the right level to pitch it. Kathy Sierra—from Creating Passionate Users—has a very interesting article entitled Reference vs. learning: pick ONE. There she explains that the author has to decide if he is writing a reference book or a learning one. There is no chance of doing both in a single volume, so one has to choose from the beginning to use the appropriate structure and language.

Writing tools

After deciding between reference and learning, the following question is how to write the notes? If I am writing short documents I normally use a word processor (even MSWord will do). However, when I work in large projects—particularly with lots of mathematics—I try to use something like L^AT_EX. If there are no maths I may use a Wiki like PmWiki. This choice does not matter at all when working solo, but what happens when I try to collaborate with other authors (a pretty common situation)? We keep sending MSWord files back and forth, but there is always a lag, because there is no way of simultaneously work in documents and proper versioning in Word. Or we… well, there is no alternative because my collaborators are no L^AT_EX or Wiki savvy.

While looking at Kathy’s article I followed a link to Jesse James Garret’s site, which in turn pointed to O’Reilly Radar, which linked to Writely’s site. Writely and Writeboard are two examples¹ of ‘web-based collaborative writing tools’. Both tools play with the same concept: let’s adapt the Wiki idea and have text documents in a web server, which can be edited by several people using some sort of markup simpler than HTML. Many wikis already have versioning systems embedded. Throw permissions on top of the mix, a simple interface, a bunch of buzzwords (say AJAX and Web 2.0) and we have the new ‘web-based collaborative writing tools’.

I have tried both systems and I feel much more comfortable with Writely. Why? Writeboard requires using the Textile markup for formatting the document. I love Textile—I am using it for writing this log entry—but my colleagues are unfamiliar with it. However, they know how to use the icons in Writely, because they use similar icons in MSWord. In addition, I can not keep a list of documents in Writeboard unless I get an account in Backpack, which is an extra hassle. Finally, it is possible to import MSWord files into Writely, which is a bonus.

I will write the notes for the first eight lectures in Writely and see if I can get some friends to make comments and modify the notes. My own little experiment, if you want.

¹ I also came across Zoho Writer, but I have not had a chance to test it.

I have been thinking about this for a while and I have decided to move most pages on this site to a Wiki¹. I do not mean the Quantum Forest weblog, which lives happily in Textpattern, but the rest of the ‘static’ (not generated ‘on the fly’) pages.

I would really like to have a system that combines wikiness and weblogging, but I have not found a nice implementation. I would really like to have a Wiki that implements Textile, but I as far as I know only Instiki (a Wiki implemented in Ruby) does it at the moment, and I do not have the privileges to install it in my system. Thus, soon I will be moving all static pages to PmWiki.

Supposedly, the biggest problem would be to transform all the HTML to PmWiki markup, but I found the HTML::Wiki Converter, which does a good job moving HTML not only to PmWiki, but to:

• DocuWiki
• Kwiki
• MediaWiki
• Moin Moin
• PHP Wiki and
• UseMod

The other issue is to redirect the old pages to the new ones, without risking losing Google (and other search engines) rankings. I will follow Google’s recommendation and use a 301 redirect, as explained here.

On top of using a ‘plain PmWiki’ implementation, I am thinking of using ‘clean URLs’ (which require these changes), user authentication (with this plugin), and extended markup (with this other plugin). The latter until I can get Textile working in PmWiki.

I will not be posting much here until I finish creating the templates, transforming content and redirecting the old pages to the new ones.

¹ A previous post considers moving only the ASReml Cookbook, but I guess it is better to move the whole site now.

PS. 2005-05-31. The HTML::Wiki Converter has a few hiccups missing a few anchors inside pages and the notation for subscripts and superscripts misses the inverted commas.

PS. 2005-06-10. It took ten days for Google and Yahoo to replace the old addresses with the new ones, so the permanent redirect did work. There was a period of three days when Google was showing both the old and new addresses in search results.

Dealing with text files requires text editors (doh!) and the choice of editor is one of the typical ‘religious wars’ in computing.

The choice seems to be between ‘do it all environments’ and ‘context specific editors’. The best representative of the first class is Emacs / Xemacs, where one can edit files, read email, browse the web, run statistical software, etc. I found the possibility of editing and building (L^A)T_EX files appealing. Even more appealing is to run S+ or R code directly using ESS. However, I feel overwhelmed by the complexity of installation and configuration of XEmacs, the different key bindings and the unnatural (to me) structure of the menus. Yes, in theory I could configure XEmacs to behave as anything I would prefer; but in practice I would need so much time to learn how to do it, that I would question if it is worth it.

Another candidate in this class is Vim. I have to say that I feel much more comfortable with it: the help system is much better, key bindings are less alien and the whole environment seems to be tailored towards productivity. I found that Vim-LaTeX (an add-on environment) is quite nice and relatively easy to use. Configuring Vim also seems easier than Emacs and I have already played a bit with the _vimrc file. On the other hand, I still do not know if there is anything similar to ESS in the Vim world; I mean syntax highlighting for S+/R and submission of selected commands. I will start learning a bit more about Vim, although I will stick to the following paragraph in the short term.

In the other extreme there are editors that excel at specific tasks. The first one that comes to mind is TEXnicCenter, which is an excellent choice for writing and processing L^AT_EX files (it is an MS Windows only program, but I guess that TeXShop is a similar choice for the Macintosh). When writing R or Splus code I have been quite happy with JGR (Java program, so runs on a Mac as well) and with ConTEXT (MS Windows only) with the optional R-Script highlighter. I think that I will stick to ‘context specific editors’. I also use ConTEXT to write ASReml code, with a highlighter that I wrote a couple of years ago.

In a related matter, when writing In defense of text, I mentioned ConT_EXt, a more modern set of macros for T_EX. Although I can see the superiority of this package, in academic circles the de facto standard is L^AT_EX, so if the writing is aimed to the formal publication process, the latter would be the obvious choice.

I have been cleaning my office as part of my implementation of GTD. I have found some very old files from the time I was at my first year of university. Some files are already unreadable in my computer: they may be Wordstar 3.3 files for which there is no import filter in MS Word. In addition, I found some old plain text files (comma separated values and T_EX documents) which I was able to read with any text editor.

Thus, the question is Why not keep (almost) everything in text files? Unfortunately, at work I do coauthor many documents where:

• I write a small part of a large report.
• I write a big part of a scientific paper with other non-text prone people.
• Everything has to be in MS Word.

However, I do write course notes and long reports where I am the single author. Here I could use the power (and long shelf life) of a markup language. Initially, I thought that the best way to create text-based documents would be L^AT_EX. This macro language for T_EX stores the files as text and when processed produces beautifully typeset documents¹. However, most documents processed by L^AT_EX look the same, and modifying (or creating) styles requires a lot of fiddling around, which implies time that I do not have.

Then enters ConT_EXt, an alternative and more modern macro package still sitting on top of T_EX. Some useful links for Context:

• Context Garden, an ‘open community website for the wonderful T_EX macro package’.
• Converting from L^AT_EX to ConT_EXt.
• etexshow, a browser for ConT_EXt commands under Emacs and XEmacs.

I am still not sure if I will move to ConT_EXt, but I believe that it is worth a try.

¹ By the way, T_EXnicCenter is by far the best L^AT_EX editor under MS Windows.

PS. 2005-04-22. I discuss some text editors in a latter post.

I finally completed (and submitted to Silvae Genetica) the manuscript for ‘Genetic variation of physical and chemical wood properties of Eucalyptus globulus‘. This is not my first or last paper (it is publication 25), but it took such a long time that it deserves a special mention. The project was plagued with problems and delays that, although did not affect the final quality of the data, made data analysis and writing the manuscript a real pain in the back.

By the way, Silvae Genetica looks like a very old fashioned 1800s journal. I always associate the image of a very old German worker printing the journal in a damp basement. Nevertheless, it is almost compulsory reading for tree breeders and the publisher seems now keen to give it a facelift.

When writing papers I use either a combination of MS Word and Endnote (a reference manager) or LaTeX in its MiKTeX incarnation with TexnicCenter as a text editor. I use the latter combination for large documents, like convoluted course notes. This time I chose Word but did not have a ‘Silvae Genetica style’ for Endnote, which is necessary to format the citations in the text. I created a style that works for journal articles, books, book chapters and conference proceedings, which you can download from here.

I do not expect to see the manuscript for around three months. By then I should receive comments (I hope positive) from the referees.