Posts about kde (old posts, page 7)

2005-03-12 23:59

Developer-centric and proud of it

Ok, as everyone and its goat in the free software/open source/pet-keeping communities already has read, Eugenia [1] has been up to her old antics.

The latest shape of her discontent is, developers of GNOME (and lately of KDE, too [2]) are developer-centric instead of user-centric.

Basically, the gist is that developers have a responsability to keep in mind their users desires, and follow their wishes when deciding what way their projects should go. That these "people" are developing said software in their free time and giving it away is unimportant, she says, the user's voice shall be heard. [3]

Well, I am not a developer nowadays, at least not of shared software [4] but I still recall when I was one. And I am cranky. And I have little care about what people think about me. So...

Let me express my feelings when I read her editorial (and another one called the riddle of the sphinx, I think [5] ).


I have read, in the editorials and (specially) the clueless talkbacks, the following:

  • If you are going to do a half-assed job, don't put it on the internet

Well, dude, pretend it's not there. That works for me when I consider japanese tentacle porn, it can work for you and a mail client. If you pretend it's not there, it won't bother you. Then those who tolerate half-assed jobs can live a life of jolly fulfillment.

  • The user has invested effort on learning and using the program, so the developer owes him something.

I suppose the solution for that is not allowing people to use the software unless they accept that it is provided as-is... oh, wait... that's already true. Never mind.

  • That since GNOME (or KDE) claim to provide useful and user-friendly software, by failing to follow the user's lead we are engaging in some sort of false advertising.

Coca Cola promised me "Everything is better with Coca Cola". Then I tried pouring a can into my computer, hoping it would make it faster, or somehow better in some way. I have to tell you, those bastards won't know what hit'em!

  • Posting bug descriptions in bugzilla is too much work.

Well, try fixing one someday, missy! That's WAY harder. You are saying that using a little energy is too much work, to people who spend hundred times more as a consequence.

It's like complaining that planting seeds is too much work, when you later get a harvest. Well, if you think, so, drop agriculture.

  • That the development is somehow skewed towards developers, not users.

Well, to that I have many answers. Let me share a few.

  1. So? You say that as if it's a bad thing.
  2. Well, that's good for the developers.
  3. What magic potion did I have to drink to become a developer? Developers are just users who are way too involved in the project.
  4. As a lawyer once said, since phoenicians invented money, the question "what can I do for you?" has a single answer. Well, think about it.
  5. No, it isn't.

Finally, let me tell you the depressing part. In a comment, Eugenia said she dropped GNOME for WindowMaker. Let me show you what she said:

RE: So Eugenia refuses to use GNOME? By Eugenia (IP: - Posted on 2005-03-10 20:32:31

I moved back to WindowMaker, yes. At least I don't expect anything more from it, because it is not a prominent environment and I know it's a much smaller project with fewer responsibilities.

Talk about throwing the baby away with the bath water. She is simply saying that she choses to use worse software (according to her needs), because... well, I have no idea why.

Because she has low-enough expectations about WindowMaker that she won't be disappointed? That road leads to dating ex-convicts, you know.

In the end, well, I have no idea what people think. I simply don't understand the world. I mean, I intensely dislike GNOME, and even a few people involved in it, but it's just code. And code that is worth millions of dollars. And it's there, dudes! It's like finding a stack of bullion in your doorstep, and not only that, but on every doorstep [6].

Why don't people simply look up and thank the invisible dwarves who must have toiled in darkness for years in order to rip it off the earth? [7]

No, what we get instead is a bunch of jewelers asking for finer craftsmanship in the goods, or else they are just gonna buy some bullion from someone else.

You know what? Developer-centric is good. In an absolute, objective, measurable way. Because it's developer-centric you get developers. developers make code. Code brings users.

If all you attract are users because being a developer is a pain the butt, you have no developers, thus no code, thus no users.

What do you think is better for users in the long term?

So, say it loud, I'm developer-centric and I'm proud. At least this old hack is.

[1] of OSNEWS.COM fame. You decide if that's a good or bad thing.

[2] A new post in her site. I suppose she has to feed the hit-monster.

[3] That being published in a big site that lacks threaded replies in 2005, and a woman who, when she gets a complain about it, said she was doing it in her spare time, so stop bothering her. No, I don't get it either.

[4] Isn't that much nicer-sounding than open source and much less annoying and pompous than free software?

[5] Which is, I think, #4 in the list of "ten titles you should never use in an opinion piece". #1 is What I did last summer, but I digress...

[6] And by some economic miracle, it hasn't caused yet a global depression in the value of bullion.

[7] Well, at least we work in comfortable chairs.

2005-03-07 23:59

Ok, so not **everyone** read my calculator rant

First of all, a note: I intend all this post as encouragement to Emiliano, the author of Kalcoolus.

You see, he turned bc into a on-screen keyboard thing . I don't like that UI much .

I can't post comments on kde-look (forgot my id, or maybe never registered), and I am not going to register just for this, but it's interesting to see how the same idea is recycled over and over and over.

This app seems to be a conflation of the following:

  • Writing frontends is good, because you don't need to write the hard part
  • A GUI calculator has to look like the real thing, in the name of usability

Well, the bad news is: writing frontends is fragile. What you want are libraries. And if GUIs had to look like the real thing, the UI to Skype would be a numpad, the UI to Amarok would look like an Ipod, and the UI to KWord would look different .

So, please, Emiliano, put a CLI in it. Pretty please? :-)

2004-12-22 18:58

Games using PyQt

As some may remember, a long time ago I started writing a Bejeweled! lookalike using PyQt.

While that game is actually playable, it mostly showed me I had no idea how to write a decent game ;-)

So, I have been for a couple of weeks trying to create a high level package for creating games, specifically board games (yes, I classify Bejeweled as a board game).

Well, it's working. I am writing a few libraries, and I am writing three games using it.

  • Scroogle: a dumb boggle/scrabble mixture
  • Nameless1: The bejeweled clone
  • Nameless2: A crafty-compatible chessboard

Since I have abstracted stuff like pieces, squares, and such, these games are quite small in size, which means easy to implement!

For example, scroogle can keep score, use a dictionary for word validation, replace used tiles with new ones, so it's basically complete.

It's under 200 lines of code.

The chess game is, again, almost complete, except for stuff like saving games, or board edition, but you can play ( it can't tell if you win or lose and has no timer, though ;-)

It's 210 lines of code (the crafty interface is 70).

Why am I mentioning this? Well, I think that if I ever make this good enough to release it, developing simple games for KDE will become much easier.

And those games will be a 80KB download, including art.

That would be good, wouldn't it? But...

... if you are thinking of writing a game, I need your help. I want you to try using this, so I can figure out if it's good or how it should be changed!

Please contact me if you know a little PyQt and want to write games!

2004-11-26 14:21

I remember worse!

Funny article at GlitchNYC about looking at KDE 2.0 today.

But hey, I wrote a theming engine for Qt 1.x, so I'vee seen, remember, and even mucked up with worse stuff :-)

2004-11-25 10:24

Interview at KDE Hispano

Late, but linked: I was interviewed by one of the nice guys at KDE hispano, something like Tink´s old interviews. In spanish.

2004-11-15 02:20

On speaking and slides

As I mentioned before, I was at the 3rd CafeLUG event on friday and saturday.

It was a pretty large event, roughly 1300 pople, 63 conferences.

I had two of my own, and here's how they went:

Friday: The PyQt/PyKDE one. Audience: about 50 sitting, 15 standing. I had no time whatsoever to prepare this. But I was luky enough to have someone (thanks Cristian!) lend me a notebook, and so I spent the previous conference (pf/OpenBSD) working on mine :-)

I kept it on a low level, nothing too complicated, since I only had 60 minutes, but the audience seems to have liked it. Or at least they found my attempts at explanation funny.

Anyway, noone was disgusted enough to tell me to my face, and I had a general feeling of it having been good.

I was able to show a quick do-nothing app in 20 lines of code, and show signals and slots, and really really push the idea of interpreted languages as better for hobby programming (which I think is the right path). Not too many questions right then, but about a dozen kids attacked me with questions on the hall after leaving.

Saturday: My generic KDE-for-beginners speech. Audience: 100 sitting, 10 standing. I was a bit too heavy on the philosophical, but it went down nicely. Several of the guys from friday's PyQt session were there, so more evidence it didn't suck too badly.

It was fun, people were surprised by Ogg-ripping from Konqueror as usual, some fun, several questions.

All in all, a nice couple of sessions, and I think I did well, which encourages me to do it more often. Who knows :-)

2004-11-10 15:16

What linux needs on the desktop is...

... oh, screw it, why not say what it doesn't need? It's better for my blood pressure.

  • It doesn't need Be refugees whining about how BeOS was so awesome and Linux desktops suck. You know what, guys? If BeOS was so cool, and Linux sucks so much, how come BeOS is dead? Now put a sock in it and go practice your OS-necrophilia.
  • It doesn't need windows users saying "it's not like windows, it sucks". If Linux was just like windows, you would say that there is no reason to switch, so put a sock in it, because you can't have it both ways.
  • It doesn't need Mac users saying how there's no point working on it because OSX is already as good a unix desktop as there's gonna be, so why bother?. Put a (designer) sock on it, because in order to use OSX you have to buy an ungodly expensive (except the 12" ibook, that's kinda cheap and cute) computer that provides only adequate performance, and I am a cheap bastard.
  • It doesn't need old unix gurus explaining how there's no need to work on KDE or GNOME because Window Maker is lighter, and all that's needed is a way to keep their xterms sorted. Put a sock in it, because xterm sucks, Window Maker is featureless, and you can't do anything useful for a regular user with just xterms and Window Maker, so you are comparing apples to apple orchards.

Most of all, it doesn't need any "What linux needs on the desktop" editorials.

On other news, I will be speaking about Linux on the desktop at the UADE in Buenos Aires, Argentina, this saturday, and about PyQt/PyKDE programming this friday.

More information at the CafeLUG site. If you read it here and mention my blog, I will pay you a beer (offers limited to a maximum stock of three beers).

2004-10-27 23:18

konqueror/embedded hacking continues

I have roughly 95% of it building with kdelibs 3.3.1, with the patches comitted up to now.

I have some weird problem with khtml_part where it doesn't see a function that's defined right there, but I am starting to get dizzy.

It's not very difficult coding (mostly cut&paste) but it's a large piece of it and this is a slooow computer :-)

2004-10-26 02:47

Konq-embed hacking progress

Ok, I have done a little about that.

Turns out I was patching it for HEAD, not 3_3_1, so I fixed that, and now the available patches work for both versions.

Of course that doesn't mean either one actually builds, only that the stuff that needed to be patched for 3.2 is now patched for newer versions as well.

The biggest problem with this piece of code is that in order to do it nicely, I have to test every change I make agains sources prepared from 3 different KDE releases (or rather, about 5 or 6, but I will not do that), so that when I touch one of the dropin/ classes to provide a 3.3 interface, I don break the 3.2 kdesrc/ build.

And I am oing it in a pretty slow box, so each build takes a couple of hours.

Ergo, I am doing it wrong. I am going to hack against 3.3.1 until it works, then 3.2 until it works, then 3.3.89.

Then all around again, until all three work without trouble. It's not going to be too easy.

To make matters worse, I realized that I don't need it to work with 3.3 for what I wanted, since a 3.2-based KHTML is already good enough.

Oh, well, at least I am hacking in public CVS again. Hopefully I won't break anything too badly.

2004-10-22 14:31

Some KDE-related hacking

Since porting HTML display engines from one place to another is all the rage these days, I did a little hacking on one of the oldest forms of the art:

KDE's kdenox module contains konq-embed, which is, among other things, a port of KHTML to almost-pure-Qt (I think).

So, I am hacking to make that really-pure-Qt.

So, it's basically porting KHTML from KDE to Qt.

Why? Because I want a decent HTML widget for PyQt that doesn't require kdelibs. So people using PyQt/Win32 can use it, for instance.

I don't know if it will work (hacking C++ code is almost painful to me nowadays).

Contents © 2000-2019 Roberto Alsina