Older blog entries for wspace (starting at number 20)

How was your holiday? Hope it was as good as mine! We stayed home, no going to faraway places this year, just enjoying the garden and home.

A friend claimed it would not be possible to do something like a roguelike via CGI. So I did a fun little project during holidays:

http://wspace.sourceforge.net/cgi-bin/webrogue/play.py?name=TestMe&pw=test&action=rest

You can get the source and some notes and a Tkinter GUI for it at

http://sourceforge.net/project/showfiles.php?group_id=72532&release_id=181250

Biggest problem was not the software design but getting CGI to work. The permission issues drove me nuts. It turned out that after creating a new db you have to wait a while before you can use it. Geez.

Unfortunately I am not a webserver wizard so my friend was right it is too slow. But it *is* fast enough with Apache running on my local machine.

If I can find somebody who can improve the performance (perhaps host it?), then it would be fun to add multiplayer and other things. The architecture is simple so adding stuff is not difficult.

There are so much people blogging now that I started to use an RSS aggregator to check what's new. I then moved all the blog URIs that supplied RSS to another subdirectory in my bookmarks (called favorites in Internet Explorer). After a while I realized that I don't use these bookmarks anymore since I read their blog via the RSS aggregator (I use effnews). I only use blog bookmarks for blogs that don't have RSS (or broken RSS :-) And I think more people are doing it this way.

So, if you have a blog, and you want to stay in people's bookmarks, don't RSS.

Heh.

Beta Education in The Netherlands and other small Western Europe countries

In my Dutch newspaper: "Multinationals complain about not enough beta scientists". Or something like that, difficult to translate. In newspaper headers they put as little words as possible, making it very hard to translate it to another language with as little words.

I don't know very much about all tech sciences, but I do know something about the computing science situation here in The Netherlands, because of my job. The situation with computing science here is worrying indeed. There are not much students doing CS at the university, it should be at least about 2 times more, year in year out. Surrounding big countries like Germany and France have twice as many students (in percentage) and they still could use some more. Why is the number of students choosing computing science here so small? Just because it is difficult? That could be the case but that would be very bad, it would mean the students are just too lazy, which does not correspond with my optimistic view.

How about prospects for the students? I feel this is the reason. If you get your M.Sc. in CS here and find a job then you will make a lot less money than if you have a M.Sc. in one of the management "sciences" or economy or even psychology, so called "gamma" sciences. It doesn't really matter what your major is. Just get a M.Sc., and as long as it is not too technical you can get a manager or salesman job and get paid a lot more. If however your education is beta, you will never have more than about 60.000 euros a year to spend, no matter how good you are. Unless you switch to a management job, showing "exceptional communication skills" and other "skills" like that. And 60k is really exceptional. The average is about 30k. Try to support a family with 30k a year. For managers and salesmen their "salary" is only part of their income, they often make more money with under-the-table-money, stock options and other extras. So the difference in expected income between beta and gamma is even much bigger than outlined here.

So it is no wonder there are not many people choosing a beta study. And the big companies (here that is Shell, Unilever, OCE, Philips, ...) used to have no problems with it. They just get more people from India and Eastern Europe and Russia and China and other poor countries. Everybody has to speak English anyway. It takes a little more time to learn English for somebody from India, which is a totally different language zone, than for somebody speaking a germanic language already, like Dutch. But look at my writings: I feel the English sucks, in style and in grammar. Believe it or not, where I work they think my English is good. So maybe it isn't such a problem after all to hire people from other language zones. At least for beta jobs.

But now these companies are getting "worried", the article says. They are afraid they will find themselves in "an unfit knowledge ecosystem". Or however you translate that. And of course they blame the government. Yeah riight. Like a government in a small country like The Netherlands can do anything about it.

Big successful software gets developed in the USA and UK, where everybody speaks English natively. Big European projects often fail because of the bad English education of the participants. This makes communicating with your coworkers really hard and often crazy. You should see the emails I get from people in France or Italy. I'd prefer them to write in their own language and then I'd use a dictionary to understand what they write. And what they write won't look so silly so it would be easier to take it serious. As it is now, you *know* they use the wrong word so a dictionary doesn't help. You have to guess.
Come to think of it, is there even one big successful European software project?

Attracting people to some kind of education is subject to (job) market workings. Like many things in this society. I think the solution to the looming future undereducation in beta science here is simple to solve. But I don't think it will happen. It's the money, stupid.

Paris

Visited Paris for some work, software design and building a prototype. Although they probably won't like it it is my intention to keep everything open source. It's a prototype anyway.

And home

I think I have some hayfever allergic reaction. Swollen inside of nose, sore throat and I feel shaky. Does not feel like high fever though. Probably caused by keeping the windows open in the hotel room. And to make matters worse, the French closed the highway leading back to my country on my trip home yesterday. So the trip took a couple of hours longer, I have seen some wayward French villages, and dozens of annoying traffic lights that make you stop for nothing.

Back home now I am sniffing and coughing but happy to see our family again, and doing something quiet and easy like watching the telly and see our man in Paris go to the Roland Garros tennis final (!), and catching up on my favorite internet readings:

• effbot

• Daily

• Ongoing

• YAFWAB (yet another Finnish Weblog about Python :-)

• Sam Ruby

• Artima

And tons of others. Ah, RSS feeds.

Family life

Our son (2 and 1/2 month old) is now able to turn his head when he is lying flat on his back. He is usually looking right and it seems independent of what there is to see. You would think that is because of the design of his bedroom, but that is not the case: his bed is on wheels and we turn it 180 degrees now and then. And it is only a couple of weeks ago that he gave us his first smile.
I think it is truely amazing to see him grow. And how. Babies can do some things already instinctively, like closing their eyes when you splatter water on their faces. And cry for food of course :-) But it is really hard to imagine the things going on in his head to become even a small kid.

Oh, I am boring again for people without kids. I know so well, happened to me all the time only about a year ago: I really got very bored when people were talking about the kids.
I will tell you what: It's FUN TO HAVE A BABY!!!

Back to computing

I finally finished a report about our project, the telecom test-case I wrote about earlier. I think I put down some good reasons to start adopting the REST architecture. In my current job I have to do a lot more writing in English instead of writing in a programming language, or writing commercial stuff in Dutch, like I used to. Writing in English is hard if you are Dutch, like me! And it has to be scientific, whatever that means. But I see so much trash in the computing science world, I can not believe my eyes. Fortunately there are gems also. That is why I switched from being a moneymaker to being a scientist. It's a severe drain of money, or how do you say that in English? I am even not a real scientist yet, I only received a M.Sc. in computing science recently. You need a Ph.D. to call yourself a real scientist here. So that is what I am working on. Pray for me.

REST

REST seems like a good idea. It is the architecture of the internet. Not that the internet was designed with it, REST has been identified much later by Roy Fielding. Some people claim that the internet is such a success because of REST. But is it? Perhaps an even better internet is possible. I like the idea of an architecture developed by evolution. But to my opinion evolution theories are not made for cases with just one species with one member.

There is a component based telecom application we need a prototype for, and REST seems just the ticket. I will give it a try. Base a design on URIs, HTTP and XML. No method calls, no CORBA or SOAP issues to deal with, just "calling" URIs and receiving XML. A big part of the design will be developing a good XML vocabulary if there is not an existing one that is good to use. So, that will make me have to learn HTTP, because I really can't stand it when I don't know much about something that seems important and I have to use it. They say one of the good things about REST is that the "HTTP gives you authentication, authorization and encryption". But I know almost nothing about how they work. They do not seem to be things programmers have to deal with much, but rather web server configuration things? I really don't want to keep some application running as a webserver, or be faced with the weird unix cron things I try to avoid, if even I have permission for such things at my ISP. I know how to do CGI programming, but would it be possible to add HTTP authentication etc. myself, or do I have to bother my ISP for that? Can anybody recommend a nice writing about HTTP? Hmm, I keep forgetting nobody reads these things, so I am going to look for a usenet group where you can ask such things.

Game programming
What a nice online risk game here. It has an active user community and many of them are having loads of fun. The guy who developed it is even making some money with it. A hackers' dream. From what I can see he uses a lot of HTML tables and he uses the *size* of the Courier font to get those numbers for armies on the right spot on the map. That could be improved upon. CSS and/or SVG is maybe better? And it is probably generated with PHP. That could also be improved upon. How about my favorite language Python.

I think every programming language could benefit from examples like that, to become more popular.

Programming these days

Currently I use Python for 3 projects at work, they are:

1. A CGI program that functions as a user interface to an XSLT transformation. So users fill in HTML forms and then some XSLT transformation is called, and the result displayed in the users' browser again. Since there is no XSLT support yet in Python, a Java transformation engine from Apache is called.

2. A CGI program that functions as user interface to a simulation of a telecom application. This will also have webservers written in Python later.

3. Research into component based programming. I use components written in Python. These components are used in 2.

Note the big CGI part in this. I would love to use something easier to work with, viz. Twisted, but I have to restrict myself to stuff available in the standard Python distributions, and then only the truly portable stuff (meaning no curses and other *n?xonly modules). For the applications the kind of protocol or other technicalities are not important.

XSLT is new for me. I have some prior experience with declarative programming (is that a good name for it?), namely with Prolog, and I used to like Prolog a lot. But my XSLT experienced is getting really spoiled because of the application: it uses XMI, that horrible XML vocabulary from OMG. Right now I am trying to find out how the hell statemachines are supposed to be encoded in XMI. The documentation (their so-called "specification") is a mess and absolutely unreadable. I will be glad when this job is done. And then, when my peace of mind has returned with respect to XML, I want to think about a fun application with XSLT.

Juri Pakaste writes about Bertrand Meyers' OO book and also mentions Dijkstra. I do have some experience with UML in industry and in academics, and I very well recognize the "trying to sell a pile of books and consulting" part Juri writes about. Amusing and amazing, it works for so many "consultants". I was there.

Did you know that the title of the famous "goto considered harmful" article was made up by the editor of a journal? Dijkstra submitted it with another title (mentioned in one of his EWDs, but I can not search those now), and everything was in a rush to finish the journal (or proceedings?) and the editor changed it to Letter to the Editor and then gave it a new title. That editor was Niklaus Wirth, of Pascal fame.

Also: Dijkstra was not a fan of OO, and did not like meaningful variables (in math:-).

More people are mentioning Dijkstra recently: Ken Arnold compares writings by Edsger Dijkstra with writings by Alan Turing with Strunk & White in his hand. Let us not forget that Dijkstra was not a native english speaker, but dutch, like me, and I can tell you that it is really hard sometimes to express yourself well in another language. Still, Ken Arnold has a(nother) writing by Dijkstra on his "topnotch" list. I think that is truly amazing for a dutchman, even if you live for a very long time in the USA.

nuncanada:         Interesting idea "Mathematics is just manipulation of strings". According to the rules of math, that is. A modern version today would maybe be "just manipulation of XML"? You'd need a MATH.DTD for that. That should be feasible. And since mathematics is often unreadable (it is not meant to be), the fact that XML is unreadable is not an extra problem in this case. And then XSLT as a math tool :-) Or am I just suffering from too much imagination again.

But I do not understand precisely what you say next. Why was formalism abandoned after what Kurt Goedel said in 1931? Because they had something better to do now, further research in this new direction, or because it was not interesting anymore (because of what Goedel said)?

Proving code

Is anybody doing something with Proof Carrying Code? Or having experience with it? Where the Code part preferably is in a modern programming language, I would like Python. I have seen some stuff with lisp, Haskell and Prolog, which seem natural choices for something like that, but these languages have too much other problems for me (no GUI modules or can't use it on my Windows laptop or no XML modules or younameit).