Switching to MikTeX

I use the cygwin based TeTeX TeX system. The problem is, it's no longer maintained and the focus is on TeXLive a newer TeX system. TeXLive has gone through several iterations and is released yearly. This year's release is sometime in July 2008 apparently.

I asked on the cygwin list whether there is going be a port of the TexLive TeX tree to cygwin. I got a short answer whether I'd like to volunteer.

ummm....Don't know anything about TeX sources and cygwin ports building and I'm not sure whether I want to.

So, I'm stuck using TeTeX and no cygwin port of the latest TeX system i.e. no upgrade path. And this is seriously showing it's age. pdfTeX version on the cygwin system is at 1.21x while the latest version is 1.40x.

Now, that leaves only one option, MikTeX.

Bit apprehensive of how this is going to pan out. Plus there's this thing of Emacs, AucTeX, cygwin and MikTeX. Windows, fake Unix, Windows port of TeX...

Combination of the good and the kludgy.

So, the plan is, I'll use the Emacs 22.1 version compiled with Auctex and MikTeX and see how it goes. If all goes, redo the same with Emacs 22.2.

Downloading the MikTeX installer and installing the additional packages now. Let's see how this works out.

OOXML and Microsoft; getting nowhere?

India has voted to NO to OOXML.

That was expected. What is surprising is how the battle is being fought. Here's one insider's account of how the votes played out and the after effects.

Reading between the lines, there seems to be a lot of quid pro quo and none of the Indian IT players come out in a better light. Of course, one is well within their rights to question the motives of the blogger too who gave his perspective of things.

However coloured the views might have been, IF there were technical demerits to the standard, then it really is a good idea to throw it out.

Microsoft has only itself to blame on this vote; none of the competition is going to help it along. Clearly, the more the NO votes, the bigger the embarrassment and consequent bureaucratic revisions and time that OOXML is going to get bogged into. The competition is also going to spread a good amount of FUD about it, right or wrong. Questions on patents, whether it is truly open, the NO votes from the academic establishments are ripe for talking points attack.

And I still don't understand why there is a need for another document standard when the ODF (Open Document Format) exists and there is ISO approval for the same.

Emacs-22.2 Released!

Saw a posting on the cygwin mailing list that Emacs-22.2 has been released.

Indeed, a quick check on the GNU website, revealed that it was done today. Unofficial cygwin builds can be found here.


Go get it.

Getting Beamer Frame Titles into Reftex TOC

I write most of my documents in LaTeX and AucTeX is indispensable in creating and generating the pdf files from LaTeX.

For presentations, one of the best available and actively maintained package is Beamer developed by Till Tantau. With its beamer templates it's possible to rustle up a fairly professional presentation in a few minutes. And if you like reading a few hundred pages of documentation, the payback is even better. :-)

RefTeX is a package for managing Labels, References, Citations and index entries with GNU Emacs.

When using the beamer package, the \frametitle{} has the slide headings and it makes sense to make that come up when C-c = is chorded. That is, when reftex-toc function is called, which builds the TOC for the document, the TOC buffer should show a number list of slide titles.

But RefTeX does not recognise \frametitle when it parses the document in its default settings.

Asking on the auctex mailing list, elicited the following solution

(setq reftex-section-levels '(("part" . 0)
                  ("chapter" . 1)
                  ("section" . 2)
                  ("subsection" . 3)
                  ("subsubsection" . 4)
                  ("paragraph" . 5)
                  ("subparagraph" . 6)
                  ("frametitle" . 7)
                  ("addchap" . -1)
                  ("addsec" . -2)))

The documentation string said that the list should be sorted, so I replicated the values instead of cons'ing and cdr'ing.


Hitting C-c = after setting it and reloading the buffer, brings up the frametitles nicely.

Using Git; more unlearning CVS

So, I trip up again when using git; Thinking that the commit works the same way as in CVS.

Of course it doesn't!

you need to do a git add and then follow it up with a git commit every time. In CVS, you need to do an add only when creating a new file. But in Git, every time the contents of the file changes you need to add before you commit.

A quick google check reveals that git tracks content not the file names. And this blogpost was even more clearer on how assuming things from other SCM versions will confuse you with Git.

To be honest, git does warn you about it and as usual, I took it to be background noise. :)

The cruft in my .emacs

It just occurred to me that I don't have a complete understanding of how I came up to speed (err,sort of) in being a bit proficient in using Emacs. I've come to realise that the advice given by Emacs developers is sound.

Do not randomly copy .emacs snippets from different users and websites. You have no idea what you copied when you're learning initially and at an intermediate stage, a little knowledge is dangerous!

My .emacs header says that it's 6 years old and I can see what a mess it is when I try to change it. I have repeated path references, lots of commented code, comments which don't make sense and some packages that I don't use at all.

If I clean up my .emacs, I'm pretty sure that I can shave off around 3 seconds of the startup time. I've also noticed that Emacs developers have made more of the customisations that were explicitly set as the defaults which means more stuff can be removed from the init file.

After backing up my .emacs I have deleted a lot of stuff and I don't miss it for over a month now. Going over the .emacs now with a bit more understanding,it's far easier to chop off code than when you're a noob. Of course, the code should not have been there in the first place.

Using Gnus to send emails through Gmail mail servers

Finally got around to updating the Emacswiki page, GnusMSMTP. which is the only page I created on the wiki so far.

You can use Gnus and msmtp to send mails through the Gmail mail servers. As mentioned in the wiki site, you might need the CVS version of Gnus for it to work. Especially if you're using the stock install of Gnus that comes with Emacs 22.1.

I use the cygwin port of msmtp and it's fairly easy to get it to work.

Follow the instructions mentioned on the wiki.

Scoring in Gnus messes up Digest mode

While reading news and mail in Gnus, I started to score authors in the Digest mails I subscribe to.

Looks like that is not working correctly, as I thought it should.

The entire digest in the Summay Buffer is highlighted instead of the articles from the specific person (which is what I score on).

Thinking this through, obviously the digest is a compendium of a number of mails with lots of headers stripped into 1 file, so the ticks seen by Gnus is on the entire digest I guess.

& you do have to hit C-d to see the articles individually anyway.

Oh well, so scoring on mail digests is not a good idea.

USENIX online conference proceedings are now freely available to everyone

Which is good news. There are some interesting papers to read; I'd recommend that you check out the best papers and the ;login: links to see some informative papers.

Wrapping my head around the distributed version control model

Reading and playing with Git, I keep comparing it with CVS and predictably end up confused. So when Kate offered to answer any questions on Git, I jumped at the chance.

One that bothered me was how the main repository was identified; Kate's explanation was good enough to clear my doubts.