Documents and stuff

Again a good text from the 750words experimentation. Remember that these are more like brain dumps than actually researched and thought texts:

Well, this day was almost gone until I remembered that hey, I still have the daily 750 words undone!

The school today was awesome for once. It seems I'm doing the IM-project I've babbled about everywhere for the project-classes, which means I have to really put my hours to a book. I resist the "use excel for bookkeeping of hours" - clause of our instructions though. .xsl(x) is a stupid format in Windows, and completely useless in the real world. Currently I do the bookkeeping as a series of Clojure maps, with keys like when, count of hours, and what did I do. I think the git could provide the when and what is done - values, but it is useless in keeping track of the hours used. I think there exist real apps for tracking the used time per project, but since the school is too lazy to propose anything better than Excel, I'm also too lazy to experiment with stuff that is not guaranteed to be better than Clojure.

If I am allowed rant more about the office formats, I want to express my utter dismay about the state of the document-producing software of ours. First: MS Word is... well, if the document is expected to follow formal standard, then it is a nice tool, although producing software with strict, formal standard is dumb. For producing somewhat informal text that doesn't look like it is done by agitated baboon it is horrible. For that I use Emacs. Today I translated and edited the first IM-blogpost into a "starting report of the project" by writing finnish into an org file in my blogging folder, and then let the Emacs export it into something that doesn't yield "WTF is this format?" - responses. I hoped the odt would be enough, since at least in the Emacs of my mac the creation of pdfs is somewhat challenged, and producing docx:s from anything usable is not a walk in a park. Of course the response was "can't you... export it to pdf or something from that linux of yours?"!

Yeah, hold on a second, I'll just brick this macbook with a fresh install of fedora or something...

I spent ten minutes trying to get the latex->pdf - exportation working, and after failing that, opened the exported odt in the Libreoffice Writer, the only thing worse in producing text documents than Word, and exported the document into pdf.

I think the project document was accepted, and now I'm rewarded with credits for working on this thing! I have yet to check the official site where these projects are hosted, but at least I received no complaints as a reply to my email where I attached the report. I was also told the school could give me a hypothetical server in case the Heroku fails, but I am not really looking forward to that, since as far as I understand, they do not have too many public IP-addresses, and I am not absolutely certain I'm able to keep the security holes out of my code.

After giving my report I resumed hacking the server of this IM (which by the way needs a real name. I think the old MEsE is a stupid one, but this faceless "the IM-project" is also such.), and found out that the server doesn't correctly update the moment of the last call any time the client calls. This means that after five minutes from the logging in the client is officially timeouted with no way to reset the timeout. Let me tell you, the state will be the end of this world.

I should do designs of the client at some point. I've spoken of this project only as a better Skype without the VoIP, but that doesn't seem to be enough to convince some people about how awesome this project is, and neither does the fact that only thing I've got to show of it are some four blog posts and a screen full of Clojure. I have seen a few people that wanted to take part in this project if I switched the implementation language to something "sane", like the Java or C# we're being taught in the school, and in a general case I could do this (although assuring them the project will probably be doomed after it), but with this project I've actually tried them. The Java was lousy in the server side, and although the C# was somewhat verbose in the client side (compare XAML to s-exprs... :P), I could see me doing an alternative implementation of the client in C# sometime.