Sun is doing quite some work on GNOME and X these days. Their latest project is to create a font library for XFree86, named Stsf, that would replace Fontconfig and Xft2. But the big question is: Does the world need yet another X font library that would create more incompatibility and fragmentation? Update: Sun proposes a new direction for both X.org and XFree86, keep reading for more.One of the reasons behind the creation of this project (but not the only reason) is the fact that XFT2 requires the XRender XFree86 extension in order to run accelerated, while Sun Solaris’ proprietary X11 and other non-XFree86 X11 implementations do not support this extension by default. So Sun got hard at work to create a new font library that would offer nice AA fonts on Java, Star/OpenOffice.org and Gnome/GTK+ applications.
However, many Gnome and GTK+ developers (and we would imagine KDE devs as well. Sun did not invite KDE to any talks when obviously that would be for the best interest of the Unix community to include them in a discussion that shapes the future of technologies that are directly involved in) are not happy with these new developments, as it would re-invent the wheel for one more time, slowing down the race to the desktop market and losing precious time. Applications would have to be modified in order to take advantage of STSF, and while there is a “wrapper/bridge” for fontconfig/xft2 apps, by using this bridge you gain nothing of STSF’s extra abilities.
“We’ve just finished spending two years moving millions of lines of code *that Sun has just shipped* over to fontconfig/Xft/Pango. Do we want to do that again?” Havoc Pennington of Gnome and Red Hat said.
Sun published a PDF on their Stsf site, comparing side by side xft2 and STSF.
“It is sometimes incorrect, often misleading, and shows a very superficial understanding of Xft. In summary, I think this paper should be removed from the authors’ site and significantly improved before it is published again” Juliusz Chroboczek wrote, the creator of Xfsft, an older font library or X.
“The document is missing the point that fontconfig/Xft2 are already widely deployed for a year or so, and STSF is still in part not implemented. This means the two technologies are not on a level playing field. For STSF to be a good idea, you have to not only show that it is better in some absolute sense, but also that
it is better enough to deprecate fontconfig/Xft2 only a year after deploying them and redo massive amounts of work. You also still have to show the server-side stuff working with good performance and
real-life significant memory savings.” Havoc Pennington replied later on a very interesting and constructive email to the list.
You can follow the discussion in the XFree Forum mailing list.
Update: Sun proposes a new direction for both X.org and XFree86 and asks for discussion upon these new points. Excert from “X.org & XFree86 relationship”:
“X.org provides funding and resources as it sees fit for activites to maintain the XFree86 software. This includes, but is not limited to, development of new features, documentation of existing interfaces, testing and maintenance activities.
Upon the release of X11R6.7, XFree86 shall integrate the X11R6.7 code into their code management system as the base of the “X.org Standards Sample Implementation” (SSI) branch.”
Havoc Pennington replied to this.
Sun has limited power to rule the community. If GNOME and KDE on Linux and FreeBSD dosen’t use Stsf, nobody will write programms for it, and it will remain as an alternative for Solaris programmers, which may result in creatin less ports for this platform.
If you want to create “more incompatibility and fragmentation” you have to have people who will follow you.
I hope that the community will not follow Sun.
I am getting fed up with this whole ‘race for the desktop’ thing. Linux, the kernel, designed to be a platform for GNU, in order to create a free UNIX clone. This has been achieved! Well done! Improve what we have and, yes, Sun can do whatever because the beauty of the Linux scene, is that there is choice.
I like to start is a minimal debian system and just add the packages I would like, and I love all this choice available. I went with KDE and ignored Gnome and only use KDE packages.
If you want to create a new desktop OS, don’t try and bring in everything available for Linux in some huge distro. Choose a kernel, choose a desktop environment (or write your own) and the apps.
It seems that whenever a weakness in Linux is discovered or pointed out (to death), at *least* four different solutions pop up. They almost never work with each other, and there never seems to be any difference to the end user. Why don’t these companies just work together to get a standard solution that may not be perfect, but at least works. Why does everything have to be so complicated?
>Linux, the kernel, designed to be a platform for GNU, in order to create a free UNIX clone
I didn’t see the word “Linux” in my article.
Oh, and do you think we don’t know what “Linux” is? But most of the time these days, Linux is meant as the “Linux platform”, not just the kernel.
>If you want to create a new desktop OS, blah blah blah blah
Well, this is what Mandrake, SuSE, Lycoris, Lindows, Xandros even Red Hat are trying to do. In other words, there is no reason for you to get fed up about the “desktop race”. There IS a desktop race happening right now from the companies mentioned above, but not from people who use Slackware and Debian (which you might be a user of).
As I said in the header, I think the good way to go is to check what sun has to offer, as any other proposal from anyone else.
If it’s really better (and no dependance from X11 is really really appreciated, you should see server applications problems due to libraries depending on X) and the code is downcompatible (they are promising even binary compatibility with Xft2 lib) and also free. What do the community have to lose? I don’t see any problem here.
If it’s better and it has all needed requirements, I think it should be adopted.
I appreciate that. All I can see is these huge bloated distros where the companies are trying to cram as much Linux software available into it.
If Sun wants to create another font library, then great! That is up to them. It is not anyones place to tell them otherwise. Redhat / Suse and whoever just should ignore it.
In my ideal world, people would take what they liked from Open Source… modified it (breaking binary compatability if need be) and create something integrated, stable and unique. Not thousands of packages being forced to work together by tying it all up with gaffa tape.
::What do the community have to lose? I don’t see any problem here.
The problem is “yet another font server”. What if in 2-3 years you get this very nifty font effect/feature that only happens on sfst but not on the (old by then) xft2 and apps haven’t migrated or don’t want to update to stsf? What you get then is inconsistencies. Stuff like that slow down the development of KDE and Gnome and their apps. If you want new features, FIX Fontconfig/xft2, DON’T re-invent the wheel.
No more fragmentation, please! I want my X desktop and my apps to be compatible with newer versions or other X desktops and apps. I want a COHERANT desktop!
Eugenia, where did you get the information that STSF was developed because Xft required the Xrender extension to run fast? I don’t think that’s the case. Actually I also think Sun has Xrender implemented in their servers. I believe the people from Xi Graphics also had problems with Xft (and even with Xrender..).
Um, because the whole point of Linux development is that the competitive environment fosters either (a) evolution, or (b) specialization. Don’t like the system? Then get out. Unfortunately, you have nowhere to go, because the real world works that way as well. Yes, this means that Linux will be a long time in coming (if ever) to the average user’s desktop. This also means that (a) The solutions that survive are absolutely top quality, and (b) Linux fills to specialize niches like nothing else. Y’all gotta realize that in the grand scheme of things, the desktop market isn’t everything. In between the server market, the corporate desktop, and specialized embedded markets, Linux has more than enough market to survive and prosper. In that time, it can let the evolutionary process create the absolute best technologies. Further, as user become more and more knowledgable and comfortable with computers, Linux becomes even more attractive.
PS> For the non-believers, think about these: the competing Linux VMs have resulted in a VM that is both fast, and scalable from embedded devices to NUMA systems. The WinXP VM is just adequate. Nothing better is needed, so no innovation occurs. No innovation occurs, so the system remains just adequate, and unsuitable for anything beyond a narrowly defined problem domain. Then think about the competing Linux filesystems: they’ve led to dramatically new approaches to filesystem design, like ReiserFS.
It is not just about Linux. This has impact to all Unices that use XFree or X11.
>Eugenia, where did you get the information that STSF was developed because….
Read it in the PDF file, under “Portability and Interoperability.” In the article, I state that this is ONE of the reasons, not “the” reason.
just amazing, wow
I read that part of Xft2 and STSF comparison and it does
_note_ state the Xrender requirement as a reason at all. The only thing it says is that STSF without XST performs about the same as Xft2 without Xrender. Thus in your article you mention only one reason that for as far as I can see is incorrect.
“One of the reasons behind the creation of this project is the fact that XFT2 requires the XRender XFree86 extension in order to run fast enough, while Sun Solaris’ proprierty X11 and other non-XFree86 X11 implementations do not support this extension.”
Incorrect. XRender is supported by OpenWindows as included with Solaris 9 12/02
From http://docs.sun.com/db/doc/816-7173/6md6rliqa?a=view
What’s New in the Solaris 9 12/02 Operating Environment
Xrender Extension
The new Xrender feature increases performance for applications, such as the StarOfficeTM software suite, that run on the Solaris operating environment. The Xrender feature provides a modern appearance for these applications. Xrender uses hardware processing for alpha-blending and transparency effects.
But what I much prefer is not throw everything guys at TrollTech, Red Hat, etc. did merely because of some technical reasons Sun seems to care about. I’m not saying this Sun thing isn’t good or isn’t bad – I don’t know. But XFT2 have been deployed for at least a year already, and is being used more and more. Sun was slower to market, I don’t think it is fair to drop Xft2 for their’s precived memory and resources enchancements.
Wouldn’t it be better if Sun implemented XRender themselves instead of reinventing the whole wheel? Just my opinion..
Presumably Sun will use it themselves, so if linux (and *BSD) used it it would actually promote interoperability between unixes right?
just my 0.02€
Both solutions XFT and Stsf are open source, i don’t think anyone looses by there being a bit of competiton. SUN are playing smart. Remember back in the 1980’s they tried to push NeWs as the windowing system for UNIX but cause it wasn’t Open Source like X it failed. Which from what i’ve read was a pity cause it sounds like it was a better windowing system. Anyway it’s SUN’s money and they can spend it on whatever they want
“I appreciate that. All I can see is these huge bloated distros where the companies are trying to cram as much Linux software available into it.”
And the reason for this being so is because you only try to see what you want. At least Lycoris, Lindows, Xandros only offer you one app of a sort.
At least stay consistent with what you said in the first place – that everything is about choice. Now, when you get a million apps on 7 CDs this does not mean that you have to install + use them all, this is the choice you were talking about. Re-think that.
It takes ten minutes of web research to pick the environment you want. Another ten to look up the most popular apps for your particular needs. Distros come with thousands of apps because (a) this is not Windows, and software distribution is nice and centralized, and (b) to accomodate people who don’t make the same choices in the above-mentioned procedure as you.
Top 5 Great Ideas:
5) One standard car. Something cross between a Ford Ranger and a Jaguar XK8.
4) One standard pair of jeans. Blue, lose-fit, stonewashed. Size 31-30. Screw people who aren’t the same size as me.
3) One standard television network. Would broadcast only government propoganda and advertisements.
2) One standard building. None of this incoherent “architecture” stuff.
1) One standard person. Wait. Never mind. We already have that. It’s called the “MTV generation.”
If i am correct, STFT will improve font rendering for StarOffice and JAVA.
Now 2 million (java)programmers are out there, and almost none of them can help in kernel, linux tools, or desktop apps.
If STFT will help java on linux, or java on GNOME or watever, and 0.001% out of 2 mil. java programmers help e.g. GNOME, then you are going to see real competition for KDE.
There’s a series of slides up which show the work that Ximian have being doing on OpenOffice, better fonts gtk+ lool, integrated printing and email (evoultion)
go to
http://www.gnome.org/~michael/XimianOOo/img0.html
(Eugenia, please spellcheck, makes your articles look a bit cheap ie proprietary instead of proprierty)
I don’t understand why people have lost the great importance that backwards compatibility has to any improvement over any software. People write new libraries that break programs dependant on old ones — what good is doing this? It may help lazy programmers get out some new code but it does nothing to help compatibility issues at all.
The lack of backward compatibility is one of the main problems in Linux. It really holds many programmers from developing for this platform.
…she’s too good for proper English.
Incidentally, Eugenia, last I checked it was your opinion that Linux didn’t have a chance on the desktop at all. Seems interesting that you’d spend so much time talking about it, in that case.
(The problem is, many people define “having a chance on the desktop” as “becoming a monopoly”. Why is that necessary? Hasn’t providing an alternative to tens of thousands of computer users already proven GNU/Linux’ worth as a desktop solution? Isn’t it improving constantly?)
He’s right. Do you cross only one style of bridge?
After seeing the screenshots, especially the font viewer, I’m convinced they should continue. Just compare these:
http://gfontview.sourceforge.net/gfontview.png
http://stsf.sourceforge.net/images/xstshowfont1.gif
Is there a kde font viewer?
I hope they push development hard on stsf and all this new X stuff. I think they should check out opentracker. Whoever finally gets that out to the masses will be remembered forever. If blueeyed comes out as soon as people are talking about, everyone may just forget X.
Anyway, I’m glad Sun is around. They’re big and clumsy, but they keep trying, and are a better friend to open source than that 800 lb. gorilla. 🙂
After seeing the screenshots, especially the font viewer, I’m convinced they should continue. Just compare these:
http://gfontview.sourceforge.net/gfontview.png
From what century did you dig that screenshot from?
Wake up and check this http://www.daa.com.au/~james/software/fontilus/fontilus-font-viewer…
“I appreciate that. All I can see is these huge bloated distros where the companies are trying to cram as much Linux software available into it.”
That is surely because downloading and installing new programs from the Internet is much less convenient than on other platforms?
When installing or updating software is as easy and simple as it is in AmigaOS, then a lean Distro will make sense.
Apparently, somebody have forgotten their “Choice is good” poster at home and woke up to a smell of coffee they didn’t like.
1. stsf also encapsualtes layout of internationalised text, ala pango.
2. stsf also has an xft/fontconfig compatability layer
3. For the particular style of thin client sun are working with stsf makes alot of sense, if pango/xft/fontconfig can’t fit thier model then sun would be daft to use it.
The arrogance of the gnome people here is palpable. The fact that sun are developing stsf in the open is a remarkable thing, that fact that redhat, who’s view of the linux desktop (inline with many views expressed regularly on osnews) is not one that many of us share, seem to feel that they solved this whole text thing, and so sun shouldnt be bothering, is laughable.
I’ve been a gnome advocate for years, but this kind of attitude is really putting me off. The fact that both gnome and KDE ship with an unuseable terminal aplication (gnome-terminal with vte is terrible, as is konsole) is typical of it.. A terminal applicaton might not be important for thier target markets, but why would people continue shipping seriously broken applications? It seems to me that for alot of people working on these core gnome technologies, they cannot be working on something that they would personally be happy using, thus losing one of the fundamental advatages of Free SOftware. Personally I’ll take pragmatic design over the supposedly academic missive of HCi researchers any day.
As a unix admin, developer and user, I am pretty happy that Sun is developing with me in mind as part of thier audience, as someone that suffered vte for far too long, I get the feeling that redhat are’nt.
Let sun work on stsf and do what is good for them, and for god’s sake can’t gnome-terminal default to zvt so that some of us can atleast get some sodding work done.
STSF Looks good to me, I think that yes it should become the new standard but xft should remain to get backwards compatibility until people can adopt the new standard
I would have never guessed so many nerds are half-wits. More choice = More fragmentation. More fragmentation = Less choice. Does no one see that integration is a GOOD THING? KEEP IT SIMPLE STUPID! So many things *nix and OSS have gone to a ludicrous amount of complexity that we are stuck in this cycle of invent -> replace. At this point in time, I propose that the CORE of graphics technology on linux be modified heavily. DRI seems to be a hot topic, and why not build font support into the X11 server. Will it help? Probably not, this is always going to happen in OSS…sure its innovation but we need to innovate ALL AT ONCE. Let’s not destroy a new standard too early, or create a new standard with limitations that will just be replaced.
P.S. I love how gnome has its OWN font manager PANGO…god damnit people.
Well… say you want THIS desktop with the BEST fonts…but it doesnt support the BEST fonts because they CHOSE to not adopt the new standard… Or say you want ARSON but you use gnome and not KDE.. (from what i have seen yesterday ARSON is the best cd burning prog for linux) Im in this position, i like gnome and i CHOOSE to not use KDE…but now i cant use ARSON because of that choice. I propose a new OpenGL implementation that is better, but further fragments the market because some people choose not to adopt it….that is a joke
the xterm in the STSF screenshot is the best I’ve ever seen on linux.
more choices ??? that’s the unix way of doing things, so they can hardly beat the M$ shit even with 30 years of maturity.
I wouldn’t expect it to be fast – OOo/StarOffice is slow anyway – so at least make them look good
Xft is a nice attempt, but the AA on LCD sux
> There’s a series of slides up which show the work that Ximian have being doing on OpenOffice, better fonts gtk+ lool, integrated printing and email (evoultion)
go to
> http://www.gnome.org/~michael/XimianOOo/img0.html
Amazing!!! ^^
> There’s a series of slides up which show the work that
> Ximian have being doing on OpenOffice, better fonts gtk+
> lool, integrated printing and email (evoultion)
> go to
> http://www.gnome.org/~michael/XimianOOo/img0.html
Cool indeed! I knew Ximian were up to something. XD2 will k|ck @ss
I don’t care who does what with the fonts, so long as somebody does something. I just booted up the newest version of Knoppix (3.2) and though the fonts weren’t the worst I’ve ever seen, they still leave much to be desired (putting it nicely). I’ve never seen any of the TrueType fonts loaded up in Linux yet, but the default fonts in most distros is a real eyesore.
So in other words, from an end user perspective, I don’t care how you improve Linux’s default fonts, just so long as it gets done.
Eugena –
In reading the article it appears that X.org is siding with stsf while Xfree86 is siding with xft2? Is my reading correct?
Given that Xfree86 has kicked out the person responsible for Xft2 (keithp), I won’t be surprised if they jump on the go the stsf way.
It’s interesting though how this STSF is now being discussed seriously only when keithp has been butted from the Xfree86 core team 8^)
There are different font formats in Windows. There are dozens of different image formats in all platforms.
Fragmentation of standards is not a linux only thing. There is OpenGL and D3D and Glide. Just because there is more choice does not necessarily mean that a new default standard is less likely to appear. Despite the existence of Directfb Xfree86 is still the standard for example.
Ext3 is the default fs on many linux versions but others choose reiser and this does not cause serious trouble.
Two different standards are out there for the desktop and I do believe this is an issue. KDE and Gnome interoperability is a real issue. I think this fact makes the linux fragmentation issue bright in the eyes of many critics and I understand this. However, the existence of other desktops like XFce barely ranks a mention. The point being as long as there is a default standard then other options are not a problem for the end-user it just gives others more flexibility.
Sun wants to do their own font thing and it has a Xft2 compatibility layer. So what? Good for them. Corporations and open source both come up with competing standards. Look at the history behind VHS/Beta or PC/Mac or a million other instances where people fight over what is the standard in the corporate world and the same thing happens in the open source model.
The moral of the story is that different people have differring opinions on attaining the same goal. Whether it is in the corporate or open source world.
If I understand correctly, Xft2 and fontconfig are relatively new libraries that were supposed to ease font configuration and they “replace” xft and freetype or something like that. What is the point of replacing something that has just come out to replace something else. This is an undending “reinvent the wheel cycle”. In Open Source world, I see this a lot – developers keep replacing existing applications instead of building on top of it.
This can be due to 3 reasons:
1. Developer mentality – “I can do this better”
2. The design of the existing application sucks and the code is incomprehensible
3. The product itself sucks
I wonder which one is the most frequent reason.
Keith Packard was the lead for Xft2, fontconfig, and RENDER and his leaving XFree86 may have something to do with this.
I do not understand the inner working of these 2 products but I personally had a very difficult time figuring out how to configure and install them due to very scarce documentation.
Okay, you wrote:
* more choice = more fragmentation
* more fragmentation = less choice
Therefore, your argument is:
* more choice = less choice
It doesn’t matter what your position on whether it’s good or bad to have multiple potential font renderers for X11, that’s logically absurd. I will give you the benefit of the doubt and presume you didn’t express yourself well, since you didn’t seem to be intentionally posing a Zen koan.
(And as is my usual argument, all that’s necessary is for people to agree on a standard high-level interface. If there’s a defined way for applications to get fonts on the screen, the underlying mechanics are of no concern to users and of very little concern to application developers–only to system developers. The problem that seems to happen in most of these “future direction of Unix” discussions is that people aren’t willing to look at APIs as distinct entities from implementations.)
Read Havoc’s email to the KDE-devel-core about choice and fragmentation. Really good explanation of when you can have choice without having fragmentation. I think it is linked on one of my editorials a few days ago.
From the screenshots it looks to me like STSF is the better of the two for quality of output. Just installed Slackware 9.0 and the anti-aliased fonts on GNOME are bloody awful.
1) You can’t compare screenshots from two different sources! They use different fonts, different options, etc. Neither Xft nor stsf handle rendering. They use a back-end renderer, usually FreeType, to do that. Quality should be near identical.
2) Luckett. I don’t know where to start. You seem to have absolutely no idea what any of these words mean. For example: DRI. How is a hardware interface for OpenGL cards going to help help 2D X rendering? You do realize that X and DRI are in two different spheres entirely? As for Pango, it’s not a font management library. It’s an internationalized text layout engine. It sits on top of the font management library so all apps can do proper text layout without the hideously complex algorithms necessary to do it properly.
3) Sukla makes an excellent point. Yes, the documentation needs work. That’s the biggest issue right now. FontConfig/Xft2/FT2 have the quality and management issues pretty much nailed down.
This is probably more related to poor fonts than a bad font system. Get the MS web fonts and see if STSF still looks better. Also, understand that your versionof FreeType also probably doesn’t all of TrueType’s hinting (due to patent Apple issues).
I don’t think that STSF is even claiming to produce better looking fonts (as it may sit on top of FreeType). It could possibly do better layout (since it actually replaces Pango), but I don’t think they claim that. Most off the claimed advantages that I’ve seen are architectural and pretty much invisible to the user. The claimed advantages also mainly come into play with a heavily networked environment (thin clients, etc…).
I haven’t heard anything so far that makes STSF a clear winner over xft/fontconfig/Pango which already has wide support. I think it would be a mistake to replace xft/fontconfig/Pango right away. Are people (other than Sun) unsatisfied with xft/fontconfig/Pango?
> Okay, you wrote:
>
> * more choice = more fragmentation
> * more fragmentation = less choice
>
> Therefore, your argument is:
>
> * more choice = less choice
>
> It doesn’t matter what your position on whether it’s good
> or bad to have multiple potential font renderers for X11,
> that’s logically absurd.
Not really, there are many *types* of choices. More *low level* choices give you fewer *high level* choices.
Let’s say you want to fragment the network protocol. Your network uses Novell’s network protocol, Microsoft network protocol (aka SMB), TCP/IP, and DEC’s network protocol. Now try writing a remote shell, … you’re in a world of pain if you do. More likely than not, you’ll pick one or two networks and implement the others “Real soon now” (i.e. only if forced to). I create a file transfer program under the same conditions and I choose a slightly different set of protocols to support. Someone else creates a remote windowing system and supports a slightly different set of protocols.
The end result is that if you want to uses these utilities together, you need to use *the least common denominator* protocol, assuming there is one.
Now assume there’s a common low level (e.g. TCP/IP and X-Windows) but fragmentation on the desktop (e.g. Gtk+, Qt, Tk, GNUstep, etc). There aren’t many problems since you can run Gtk+ apps in KDE and Qt apps in GNOME, and Tk or GNUstep apps in both. We can do this because we agree on the fundamentals. In this case, fragmentation has given us choice.
Ill check it out. And i purposefully wrote that WattsM, because this is the case in my opinion. you get more choice of one thing, but it limits your available choices of another, equally important, thing. STANDARDS is what we need…and we need GOOD standards that we wont attempt, and fail, to throw away in a year or two. The way i see it, a HUGE problem with OSS is that once a standard has been adopted, albeit by half of the developers, the standard will remain the standard in some of the applications and the others will adopt an even newer standard, or not. I would also like to mention that with all the beauties of OSS, why isn’t our source code reused more often? Isn’t that the point, and one of the main advantages we have over closed source? <- Reinventing the wheel, is stupid…and i pray that we go ahead and make a damn good road map that the major players can agree on, and everyone will be forced to follow this roadmap and adopt whatever new (and hopefully final) standards it entails. I know this cannot happen, because of personal choice of the developers, but it would be so beautiful. Basically…we need a common API. and an integrated solution. Xserver (or not) which has DIRECT ties with fonts, applications, input, hardware. Also a modular approach to font handling would be nice, is it possible to incorporate a framework in all of the toolkits (gtk, qt, wxWindows, tcl/tk (or whatever, wine even) where you can still choose what system you want and have the X server pass it on to all the applications. Hopefully this is what STSF has in store. Im really just rambling and you all know…but im frustrated with the way things are going…and might as well spam my ideas here…cuz im bored.
Thanks
Or put another way, People can eventually adapt to drastically different classes of vehicles (cars, 18-wheelers, tractors, motorcycles, minivans). Each type of vehicle requires different skills, but that’s okay.
However, but when each road has different rules (drive on right side for some roads, drive on left side on other roads, or drive which-ever way gets you ahead on other roads), the only thing that’s going to happen is traffic accidents and driving ticket.
Choice in one case is good. Choice in the other is bad and ends up restricting you to roads where you already know the rules.
that way we save a lot of time/effort spent on desktop linux.
Java works as is and we will wait till a GTK port is available for AA. True java looks crummy on linux and doesn’t interoperate with other apps but at least it works until a new port is ready.
The first ‘disagreement’ with me wasn’t that much of a disagreement; the terms ‘high level’ and ‘low level’ were being flipped around. Perhaps we should say ‘application level,’ ‘interface level’ and ‘system level.’
If there is a common standard at the interface level, then it largely doesn’t matter what’s being done on either the application level or system level. To use the TCP/IP example given, the specifics of how the TCP/IP stack are implemented aren’t material–only that the interface the stack presents to programs trying to use it is consistent.
Ket’s turn that example around a little. If I have standard commands for opening a socket, sending and receiving data over it, and closing it, it doesn’t matter whether the underlying network protocol is TCP/IP, IPX or something else–it only matters that it presents standard commands that I could use. On the other hand, if you and I both have TCP/IP stacks but mineis written in a way that programs just have to be relinked to my library–or better yet, it functions as a drop-in replacement for an existing library–while yours implements the same concepts in an utterly different way that will require prorams to be rewritten, it doesn’t matter that we’re both using the same low-level protocol. We’ve created a schism, and the chances are that mine will “win” even if yours is considerably better.
Bringing it back to this discussion, all that matters is for “font engines” to have an agreed upon interface level (for practical purposes, that includes font formats and other things that make it hairier than the network stack example, I know). If they don’t, then whichever one’s compatible with what’s most widely in use pretty much wins by default. If they do, this is a tempest in a teapot.
STSF will support multiple X implementations (XFree86, OpenWindows at least) and at the present time Xft2 does not.
STSF is backwards compatible with Xft and will therefore bring Xft support to OpenWindows.
STSF means a great deal of Sun because of this… they can deploy STSF in conjunction with the Sunray Server and have antialiased font rendering on Sunray networks.
While, on the flip side…
Xft2 is a working solution now… for XFree86 only…
Well, it’s open source development, and people do pretty much whatever the hell they want, since no one is in charge. I think we’ll see parallel development of STSF and Xft2, with no clear “winner” for quite some time.
I would assume from Sun’s vested interest in Gnome that GTK will get STSF support rather soon, though…
Keith left X etc – but I haven’t heard anything else on the topic yet?
Is there a fork around or something?
X is the biggest POS ever, I’m crossing my fingers that finally some rapid open development (and big changes) can finally make their way into X with the recent events.