Slashdot Mirror
Rasterman leaves RedHat
Well I thought I'd send this mail out now...
The short and curly:
As of friday the 28th of May I no longer work at Red Hat Software. The story is along ad will be explained sometime later - it's been brewing for months though.
I am now officially unemployed.
I am as happy as chickens in a seed farm.
As of tomorrow , June 1, 1999 I will be leaving North Carolina and moving west. I currently have no other employers.
This is all good for E and E development.
I am so glad to get out of ths creativity-stifling environemnt of RHAD LABS - away from certain people there who see E and its userbase as what I can literally quote them saying a "festering crowd".
I've tolerated this attitude towards E users for way too long. I do not envisage much future support from Red Hat for Enlightenment - they have been itching to get rid of E and will as soon as they can. I don't much care. They can do whatever they like - and I wish them all the best but I do not fit in there. E does not fit in there. They want a windows clone distribution and OS. I do not. They don't believe users really count - corporates and "partners" count and what they percieve as the "business world that wants an exact windows clone" counts.
I am not advocating changing distribtuions, but I am saying that this is onyl good for E - I will be freed up to concentrate on it and associated projects (that includes working on X and extensions to X). This also means i will be able to develop E free of GNOME. E is NOT GNOME's Window Manager - GNOME does not have one. Infact E will be workign to becoming its own desktop shell (I separate Desktop and Desktop Shell here for a reason) in its own right as time moves on - but unlike GNOME I won't make a vaporware publicity stunt out of it until there's something concrete there. E is getting on and a lot of important backend code is in place. After a few more necessary features it will start to grow into a desktop shell (a desktop shell is what I term the combination of Filemanager, Window manager and a Panel app launcher and an "applet/dock" holder). This means E is independant of whatvere desktop apps you use - you can use KDE apps, GNOME apps, GTK apps, X apps, Motif apps, CDE apps - whatever apps you like - but your desktop shell will be consistent and configurable to exactly how you want it. You alreayd knwo E's memory footprint is pretty small - especialyl compared to those of gnome and KDE (when you add the memory use of all the applets, panels, programs, wm etc. of each they add up to much much much more than E). E can absorb much of the functionality of these with very little overhead since it's already got the backend code there in E. Once the desktop shell for E is compleyte, debugged, optimised and so on E will hit 1.0 - but I'mnto setting a time limit on this - this happens when it's done and not before.
If you want to help: sit tight and stick to E - send good feedback and bug reports - We DO listen to them. Send patches to E if you want features. When the new dock applet apiu is done you can all scrutty off and write 5000 loadmeters, cd pplayers etc for E's new dock applet api (yes non square 64x64 dock apps - any size, anywhere (in the dragbar, in small windows, on the desktop istelf). all "dock applets" for E will follow the theme of the WM. This has yet to be worked on but will be - as well as adding in the iconbox again (that comes first).
Expect E to go far.
For the next week I'll be driving across the USA so dont' expect much response from me - after that I'll be moving into a new home, but therafter expect things to move along again.
I do hope we are doing things right by the majority of e users out there. You are my priority - not commerical interests, not political games, not a windows clone, not GNOME, not KDE - users come first. Those that help wiht the project get their wishes often done sooner than others becuase there aren't many working on E.
Here's to a bright future for Enlightenemnt and for all who use it.
--
--------------- Codito, ergo sum - "I code, therefore I am" --------------------
The Rasterman (Carsten Haitzler) raster@rasterman.com
Raster's Page raster@linux.com
"
Anyhow, E's look is nice and all, but gnome's outright refusal to adopt a window manager has annoyed me to no end. I change gtk, I have to change gnome separately. Gtk and gnome never look quite like they match up, and of course there's a whole new drawing layer on it.
... I simply have no link to reply to this article, and the moderator dropdowns show up inlined. kfm seems to be awful funny with forms... So anyhow I'll piggyback my reply to someone famous, hopefully it won't be moderated down :)
As for Kwm (Raster does talk about E in KDE after all), it uses Qt as its toolkit. Same idea as mwm using motif. The window menus and root menus and such, they're KDE menu objects (derived from Qt). No wheels reinvented here.
In the gnome app list, I see a gtkwm that aspires to do the same, but it appears to be forever vaporware. Could some enterprising soul who knows window managers perhaps take up this project? It's ridiculous to have a window manager represent a code fork from a desktop environment.
As for reply links
I've finally had it: until slashdot gets article moderation, I am not coming back.
actually it's not GPL anymore, it's the X license.
which is actually a lot freer than GPL.
and you'll find a lot more people than raster's name in the copyright.
eesh -ewait "copyright"
--
Geoff Harrison (http://mandrake.net)
Senior Software Engineer - VA Linux Labs (http://www.valinux.com)
Geoff "Mandrake" Harrison
Some Random UI Hacker
If I had achieved 10% of what Matthias has, I would be 10 times more ;-).
annoying than he is (hey, I may already be
Besides, you are totally missing the point of what you replied to:
Matthias is just a developer. He is an extraordinary developer. Hell, he
writes perhaps the prettiest code I've ever seen, but he is not the
equivalent of Linus/RMS/Miguel/whatever in one important aspect: he is not
"the boss".
You see, he is not humble about his code. He doesn't need to, his code is
good. He is humble about just being a coder, which is (IMHO) a lot more
important.
BTW: what he says of KWord and Abiword is actually true.
Well, this is kind of strange.. Its never good to hear this kind of stuff happen in the family. He alluded to the fact that this sort of tension had been brewing for several months, which makes me think this isnt exactly a spur of the moment decision for him.
Whatever his decision is, thats cool. It's Karsten's right to flip the bird to whoever he wants..but not without consequences. You have to admit, there are better ways to leave a company..Certainly more professional (and perhaps more mature) ways, at that.
Publically referring to your former employer as an entity which doesn't care about its user base, and prefers only "commercial interests, political games, and making a windows clone" won't exactly earn you any friends. Or a good employment reference, for that matter. It makes both Red Hat _and_ him look bad.
My only fear is that whoever his future employer will be will look past his talent and see him simply as a potential risk to the company's public image. If he decided he didn't like my company, and then turned around and slammed MY company in the press like this, I sure as hell wouldn't hire him either, talent or no talent.
If you dont like where you are, thats fine. Its cool to move on.. But dont spit in the face of people who gave you a shot in the first place. Its not just unprofessional..its also a bit childish, imho.
Bowie
PROPAGANDA
Bowie J. Poag
I've noticed a number of instances where KDE has really had a different culture than GNOME or E, which has led to a different attitude to from the project:
1. Core KDE developers *never* rip on opposing projects. They attempt to intergrate. I'm sure everyone remembers when KDE anounced Version 1.0 of GNOME promenantly on its webpage.
2. No publicity stunts. The software's done when it's done. 1.1.1 took forever to get out, but when it was out it worked really well. In fact KDE folks are currently debating whether they should try to pre-release more official stuff in order to generate interest. For example, the koffice daily snapshots won't generate the type of interest a Koffice-0.3 would.
3. No central cultish leader. Sure it's nice to have what ESR described as the "benevilent dictator", the Linus or the Larry Wall. However, neither Miguel or Rasterman fill this role particularly well. They have the dictator thing down, but not the benevilence. Sometimes, I'd say the FreeBSD/XFree/Apache model of just a bunch of developers works pretty well.
Just some observations about the way Open Source software works in different cultures.
"more stable" ?
"lighter" ?
most people who say these types of things don't know much about enlightenment. it doesn't hog memory unless you tell it to, and it certainly doesn't crash all the time like people tend to assume it does.
--
Geoff Harrison (http://mandrake.net)
Senior Software Engineer - VA Linux Labs (http://www.valinux.com)
Geoff "Mandrake" Harrison
Some Random UI Hacker
This is arrant flame-bait and not at all necessary. For a long time I used Gnome on one machine and KDE on another. After RH 6.0 came out, I chose Gnome everywhere and erased KDE. Is it so hard for KDE folks to believe that some of us prefer Gnome, and have not experienced any sort of stability problems with it? To be sure, I didn't have any stability problems with KDE either. They both crash much less than Win98, which I am forced to use for some purposes.
I went from Gnome 0.3 to 0.9, where I stayed for a long time. When Gnome 1.0 was released, I heard a lot of well-considered criticism of its stability and avoided it. I heard that it was fixed by about 1.03 and that's when I first upgraded to 1.x. I simply do _not_ have problems with crashes on my machine: Celeron 400 with 128MB RAM running RH 6.0 + Gnome (RH 5.2 + Gnome + Kernel 2.2. was just as rock stable). Netscape is the only occasional offender, and nothing that kill can't fix. BTW, I do everything from burning CDs to GIMP art to programming (Python, Java, C), etc on my main machine.
Gnome has as much a place as KDE. The competition between them has made both _much_ better in spite of the childish jabs from either side. Miguel can sometimes overdo the advocacy, but I have also heard (at second hand, admittedly) hair-raising flame-bait from core KDE developers as well.
And the key point is that Red Hat is not "pissing away" anything by supporting Gnome. In UIs, as in other things, non est disputandum de gustibus. Believe it or not, the fact that RH didn't board your favorite ship does not doom it to oblivion. Nor does the fact that other distros chose another desktop. All it proves is that there is healthy competition in the Linux landscape.
--Uche
"What thou lovest well remains, the rest is dross" -- E.P.
I was rather pleased to see how Gnome was coming along because I don't rather fancy the look of KDE, although that is what I have been using for the computer labs because it is the most complete system at this point.
Raster is correct in that E and Gnome are two different beasts. That is the problem. In giving out Red Hat 6.0 CDs to Linux newbies, I have found that they all get incredibly confused by the fact that the level of complexity of Gnome is compounded by a factor of two because of the fact that one must configure the window manager and gnome (sorta like matching your shirt and trousers).
I think that Red Hat was trying to push E towards being "Gnome's window manager" simply because that is what the most people out here would really really like to see. Gnome fills in the gaps that E leaves rather well, and truth be known, I think that a total integration would make the most sense. Insisting that Gnome be "non window manager specific" is just plain insane on their parts: it NEEDS to be or will forever have that dual configuation hell.
Despite what might be best for Red Hat, or what the most people want, absorbing E into Gnome most certainly isn't Rasterman would want. I can see why he would feel this way. People do open source software not for the money, but for the glory. There is precious little glory having your work buried into another project.
So, unfortunately, this shows one of the major weaknesses of Open Source. Because the modivation is notority, it lends itself to programmers whose egos can dictate more than what might be good for the community. Raster meantions that he is motivated by user input, but from what I have seen with the people I have tried to introduce Linux to, an integrated Gnome/E would be the most preferable path to take. While really really pretty, Enlightenment has always been the least usable window manager in any incarnation.
Admittedly, I don't know what might have gone on inside Red Hat, so I apologise to Rasterman if these comments have sounded overly critical, and certainly neither I nor anyone else should have the right to dictate the course of your life. But I must say that I am disappointed to see that we are less likely to see a more integrated Gnome/E and very disappointed that the change could not have been done in a more gentlemanly manner.
Surely this will provide a dividing line for the community and a oil tankers worth of fuel for the flame war that will follow.
There is no entry on his page that he is leaving RedHat. However, he DOES say that he is moving to California:
So, we know for sure that he is leaving North Carolina, but not that he is quitting RedHat. Can anyone verify more than this? Was his post signed?I met Carsten once, at the 1998 Linux Expo (and what a blast that was). He struck me as a highly creative mind, with a desire to create software that is innovative above all else.
RH, it seems, wants to target the mainstream user. Simplicity is their goal.
Both are great goals. But these are divergent goals. it is sad to hear of the sniping and infighting in RH. But, perhaps, this is more proof the the Linux Revolution (tm) is in full swing. This simple software can be stretched into countless directions. In essence, this CAN be a good thing for us all.
Good luck, Raster. I will keep your RH business card as a collector's item.
JL Culp
My palmtop is a 486. I care it runs stuff well. It doesn't run E too well, which I don't care about and imlib is awful on it, although better since I bitched at Raster.
The 20 seconds delay on a 486SX caused by imlib poor coding is a 4 second delay on a pentium 166 which for 10 gnome apps starting is a lot of CPU time.
People who write unjustifiably unoptimised code are not good programmers. People who write inconsistent guis are not the greatest gui designers.
There is a lot of E code that is justifiably CPU intensive. It isnt rasters fault shape extension in X11 is heavy nor that transparent window moves while they look beautiful are CPU heavy. Imlib on the other hand I don't like codewise. I use it cos it works. (That being qualification #1 for good software 8)).
Another way to think of it for the more religious warfare inclined - imlib is why KDE is a lot faster than gnome on an 8bit display lower end machine.
Red Hat is in the middle of a conversion from a small, tightly-integrated company with a strong shared vision of its beliefs and values, to a larger firm, a notable industry player, with dissenting versions of what its vision should be.
No doubt there are people at Red Hat who think that producing a Windows-clone user interface is the best way to going beyond the early adopters and penetrating the majority market. No doubt there are also people at Red Hat who think that the whole point of the exercise has been to build something different from Microsoft's offerings and that if you're just going to turn the product into a Windows-clone, why bother? No doubt there are even people at Red Hat who don't care a great deal about these issues, and just want to do their job and pick up a paycheck.
This is normal.
I'm not saying that I agree with everything Red Hat is doing; in fact, I have had serious issues with Red Hat for a couple years now. But this item is not one of them. We can only expect to hear more and more dissent from inside Red Hat, and this is good--it means they are maintaining transparency. We as users and customers want to know what's going on inside the company, and that means sometimes we will see some dirty laundry aired. Let's try to be mature about it.
-Graham
I'm not sure why there's always been an anti-biz slant to just about every one of these divorces. Sometimes, and I've been there and done that, one has to do code which productizes software.
Overall, I think RH is on the right track to making Linux (or broadly, Unix, since most Linux apps will run on other x86 Unix platforms) usable for the masses. Yes, it means "selling out" to "the corporates" -- but it's not selling out, per se. The reason why mediocrity has 90% of the users out there is because it has an incredible marketing machine (and really crafty contract writers.) Part of the RH strategy is to give the ordinary (dare I say, mediocre?) users something they're familiar with... and then be completely subversive by introducing changes that get them on the road to something more useful (UI-wise, application-wise, etc.)
While I appreciate and try to code to perfection too, there's a point at which the code has to be released and shipped. Usually companies have two sets of coders: one which creates the features and the "cool stuff" and another set which productizes. It seems to me that RHAD combined the two, pissed off one of their developers, and now we all get to hear about the bad blood.
Personally, I hope there's an opening when I finish this degree at RHAD. The important lesson is to choke down one's ego when it's appropriate.
-scooter
Why do you think that this is the truth? This is one persons view of things. How much did raster have to do with GNOME? The only think I see is his window manager specs. He most likely did not work well in the RHAD environment because it came time to get some real work done - i.e. - Red Hat Linux 6.0 needed to have a good desktop. raster has never worried about deadlines or backwards compatibility. Just look at how many times theme creaters have to rewrite their themes because of config file format changes.
Furthermore, Red Hat is doing the furthest thing from making a Windows clone. It may be true that they made the Clean theme - the most Windows looking theme - the default in Red Hat Linux 6.0. That decision HARDLY makes Red Hat Linux a windows clone.
Just by saying "unlike GNOME I won't make a vaporware publicity stunt out of it" you've done just that - in saying these things about your co-workers - your colleagues - it shows that raster just didn't fit in. It seems that he had his own agenda and just wasn't a team player in the labs.
Red Hat wants to bring Linux to the masses. Raster wants to write really cool code. These are both worthy goals, and they often overlap; but often they don't, such as when the masses want something boring.
So it makes sense for Raster to go his own way, and I don't think it's a disaster or a tragedy or a commentary on free software development; he and the company just didn't fit together well, that's all. He's variably sized and curvy, Red Hat is 64x64 and square.
People change jobs. It happens. Often it's good.
Alan
who prefers his windows rectangular and opaque, but still appreciates Raster's work
Posted by Federico Mena-Quintero:
I think you are referring to an article about RHAD Labs that appeared some months ago in the News and Observer (a local paper). It was me who made similar, but not quite the same comments.
Let me tell you the little story. The day before the interview I had spent an unfruitful afternoon trying to write a configuration file for Enlightenment, to make a theme. Mind you, I did not succeed. So the next day I was rambling about how unreadable Enlightenment's configuration files are. I was especially untactful against the Enlightenment configuration parser.
So on the next day there came this reporter to interview the RHAD Labs crowd, and he asked me what I was rambling about. I told him, and somehow he managed to mangle it into text that said that I hated Raster's code and that Enlightenment was a bad program. I don't think I said that. I said that I did not like the way it parses configuration files. I apologize to Raster if I was harsh.
I have offered Raster to write a real parser for his Enlightenment configuration files. I think it is important for the window manager such as Enlightenment to have an easy way to create themes. A configuration file with better organization than what there is now would be good. A GUI tool to do it would be even better.
Again, I am sorry. I have learned the hard way that the press likes to mangle one's thoughts to create "interesting" press about a non-existent conflict.
As Miguel said, the thing that we *do* want to replace is Imlib. Imlib's memory management is rather poor, and as such GNOME applications cannot be as efficient as they could be with respect to image loading and caching. Imlib was designed for Enlightenment's particular needs and as a libXpm replacement.
I hope this clears things up. I apologize to Raster if that newspaper article implied offensive things. And I wish him the best of luck in California, where you can find nice civilized cities.
Federico
Back when Raster started working at Redhat I was kinda wondering how it would turn out. I'm not terribly surprised to see the outcome, but a bit disappointed in Redhat. I don't use the distribution myself, but with the price raise to $80 for 6.0 (At the local Best Buy anyway), and Raster's comments here, it really seems like they may be headed towards being exactly what so many of us feared. This isn't to say that Redhat is going to become some kind of dictator, but that thier goals as a business, not just succeeding, but excessively growing, seem to be getting higher priority than than just developing cool software.
;)
I understand that making sure you have a good business model is certainly important, but when any company, or even a group of people, target to have a monopoly on the interface to linux, i.e. what a person sees, when they see it, how they interact with it, that strikes me as being rather scary. This is, imho, why Enlightenment seems such a good concept. Provide the backbone to allow any interface, but do everything behind the scenes. Don't market the default interface as the end all. Actively promote diversity to accomplish the best results. I truely hope in the future that linux doesn't simply become the underlying archetecture to propriatary interfaces. Even with opensourced code, the concept of an end-all interface with it's own agenda can be damaging. Unfortunatly, it appears that this is the road that many companies, (and individuals) desire. Hopefully the most open design in the end will win, and that the people who work for companies investing in linux, will not abuse thier positions.
Ok, I'm off my soapbox now, flame away..
You got your references wrong Dude.
It was not me trying to rewrite E, probably someone else, but not me. As far as I care, I only care about the application framework and the applications.
The window managers never quite excited me, so I doubt it was me.
The only code I want to rewrite is Imlib, because Imlib 1.xx has serious memory management issues. So we are going to base our new image code in Raph Levien's code.
Hope this clears up your confussion.
Miguel.
Interesting commentary indeed.
I wish Raster the best of luck. I'm not sure why he would care to say nasty things about Red Hat, the product, or the people there, though. I don't know anyone at Red Hat who did anything to make Red Hat an unhappy place for him.
Sometimes things just aren't a perfect fit, though, and people have to find that perfect fit. I hope raster finds his. I'm saddened at Red Hat wasn't it.
I also hope that E development continues as it has. I think it's a damned good piece of free software and hope to continue to be able to use it. I also hope it does continue to tie into GNOME nicely. No, it doesn't have to be the GNOME window manager. There can be many. But I like E and I think it fits well.
Anyway, good luck raster.
--Donnie
I'm currently trying to build a really small linux installation for 486es, and I'm getting the files from RH5.1 to do it. The CD originally came from linuxmall as a two-for-one at about $5. Red Hat is probably not the distribution I should be using, but it's the only one we got at the moment, we're real low budget
Anyone who seriously thinks that cloning Windows is strategically vital had better go investigate the Interface Hall of Shame, and the reviews of the Windows Find applet, Explorer, and the common file dialogs. These are faithfully duplicated in environments like KDE (I'm thinking of Explorer in particular, it is _very_ similar), and the agenda to clone Windows will bring more and more of these horrible, appalling errors and awkwardnesses into whichever Linux environment goes that route.
Meanwhile, I'll be messing around with largely text-oriented Window Maker implementations (and figuring out neat things to do with scripts), and Raster will presumably be constantly furthering the limits of wild and ornate window manager interface design, and we can damned well make our _own_ mistakes, thank you: we don't _have_ to make Windows' mistakes as well just to be taken seriously. I'll happily take Raster seriously- he talks like a designer, like someone willing to try something new, or make his own decisions. I hope he takes me seriously but hey, I haven't 'shipped' yet so I have to get results together before I can expect to even be noticed. At any rate, I think it's safe to say that neither of us give a damn for faithfully replicating Windows mistakes out of some misguided notion that it is expected of us
So good luck, and if there's anything I can do to help, Raster, you're welcome to it. Here, it's not much, but I am good with GFX: use any or all of my Linux graphics such as tiles and textures and backgrounds. If I can do more I will, and if my own pursuits help you out I will rejoice, just as I daresay you'd rejoice if yours help out mine.
And if Red Hat does not rejoice to see non-Red-Hat-style implementations being busily developed, if they do not rejoice to see their profitable standardization undercut by people like us, well, fsckem
- if you want windows so badly, Microsoft is happy to sell you it
- cheapbytes. Who pays the packager/distributor $80 for what is free, particularly if it isn't in turn funding the Rastermans of the world? Who'll pay Red Hat to make Linux more like Windows? Not me, I'll tell you. They are just another distribution.
Good luck, Rasterman. Hack on.Try inflicting Enlightenment on your grandmother or using imlib on a 486SX machine. There will always be a difference between end user ease and reliability and the Rastermans flair for the bizarre and incredibly flexible.
Enlightenment is a beautiful toy, if you want to do wild and wonderful things. But to a lot of people the fact that all buttons behave the same way is a feature they like.
Good luck Raster, E may not be the WM everyone uses at work, but its the one everyone uses for shows.
Alan