Omni Group is well known to most Mac users and NeXT ones back in the day. Omni Group today is a MacOSX-only company, writting high quality applications in Cocoa, apps like OmniWeb, OmniGraffle, OmniOutliner and 3D game ports. Today, we talked to its CEO, Ken Case, about the company and its products, Apple’s strategies, Safari and the future.
1. OmniGroup’s best known product is OmniWeb, a modern web browser for OSX. Recently, Apple released Safari, which seems to have already captured the interest of the OSX userbase. What market problems would you think this release from Apple will bring to your product? Is the OSX community big enough for seven browsers, especially after the Safari release?
Ken Case: OmniWeb is a browser which provides a very rich browsing experience, and is a very successful product for us despite free competition from the web industry’s giants, Microsoft and Netscape. OmniWeb’s biggest weakness has been a lack of compatibility with some web pages, and solving this by implementing newer web standards is the focus of our current efforts on OmniWeb.
Safari appears to be a great alternative to Internet Explorer as a free web browser which ships with the operating system. It seems to be
quite fast, small, and easy to use (much like the new 12″ PowerBook), and I’m very glad to see Apple basing their product on standards and
open source technologies.
I don’t really see Safari as competition to OmniWeb: they’re aiming at capturing the entry-level browser user (that currently sticks with the
bundled Internet Explorer), while we sell OmniWeb to those who really want the most efficient, powerful browsing experience possible.
Until we finish our work on newer web standards, we’re our own worst enemy: I don’t believe that other browsers are pulling any of our
customers away from OmniWeb, but that we’re pushing our own customers away from OmniWeb while that work remains incomplete. Fortunately, I think we’ll be there soon!
2. On a similar subject, recently Apple stirred up some controversy online with its release of the new version of Sherlock, which does most of what the popular shareware app Watson does. Do you think that Apple, as the producer of the OS, should also be a major application creator? Do you think that these “iApps” will create problems in the development community and limit their profits in the long run?
Ken Case: Apple is unique among mass-market computer companies in that they
provide a complete integrated solution which includes the hardware, the operating system, and a number of (bundled and unbundled) applications. This allows them to innovate in ways that involve the complete integrated package, rather than being limited to innovation within each level.
This may make them better competition than the average competitor
(because they’re hopefully providing better solutions than
average)–but I don’t see it as fundamentally different from any other
competition from any other source, as long as they make the same
information and API available to third party application developers as
they make available to their own application developers.
3. OmniWeb has its roots to NeXT. How close to NeXT do you find OSX today? Are there features (as a user and developer) that you had on NeXT but you miss on OSX? What are your favorite features on OSX and what features do you think that OSX lacks?
Ken Case: We developed applications for the NeXT platform because we found it was
an incredibly productive development environment for us: it was a
wonderful blending of UNIX with a modern object-oriented toolkit. In
that respect, Mac OS X definitely takes up where NeXT left off: we
have a more modern (and better supported) UNIX, and Cocoa is a
wonderful evolution of the NeXT toolkits. And on top of that we have a
lot of wonderful Macintosh software through Carbon and Classic.
But Mac OS X goes even further than that: we not only have UNIX, and
Cocoa, and Carbon (and the ability to leverage them all from within the
same application), but other great standards like OpenGL, and all
within a new Aqua interface which encourages application designers to
avoid unnecessary complexity within their applications without being
overly restrictive in their simplification, through dynamic elements
like sheets, drawers, and expandable interfaces.
As a developer, though, I do miss one thing from the NeXT heritage:
their Enterprise Objects Framework for OpenStep, which made developing
powerful database applications a dream.
4. Do you have plans to extend OmniWeb’s current HTML engine with better js and CSS compatibility or do you have plans to reuse another HTML engine like WebCore or Gecko? What features will the next version of OmniWeb have? Any
plans for tabbed browsing?
Ken Case: We were extending OmniWeb’s current HTML engine when Apple announced their plans for Safari, but now we’re taking a hard look at their work on KHTML to see whether that might be a good solution for OmniWeb.
We’d looked at Gecko in the past, and portions of it–like the
SpiderMonkey JavaScript engine–are currently used in OmniWeb. But the
design goals for Gecko as a whole didn’t really mesh very well with
OmniWeb: it’s a least-common-denominator operating system solution,
while OmniWeb tries to make the most of our underlying operating system.
We’ve just released OmniWeb 4.2 beta 1 this week, and (in addition to a
number of other enhancements and bug fixes) it includes a ‘Zoomed
Editor’ for entering large amounts of text into forms–it lets you
enter the text in a separate window that you can resize to see as much
or as little of text as you’d like. But 4.2 doesn’t include a new
rendering engine yet, so that remains our biggest weakness at the
moment.
For OmniWeb 5.0, we do plan to provide a solution for simultaneously
browsing multiple pages in the same window, and we’re working on some
new browsing innovations. But first things first: before improving
our award-winning interface even more, we’ve really got to fix our
rendering engine.
5. A problem I have personally experienced with all the OSX browsers on my G4 is that scrolling and especially resizing web pages on OSX can be slow and jerky (as opposed to smooth OS9). Why does this happen? What are the purely technical reasons that make scrolling quick and smooth on other platforms but are a problem on OSX for many users?
Ken Case: I think that the big limitations here are that Mac OS X’s Quartz
interface is currently software based (since no hardware exists which
supports things like rendering nice antialiased text), and that the
PowerPC chipset doesn’t have a direct equivalent to the Pentium’s
write-combining mode for video memory regions. If we could render in
hardware (as today’s 3D games do), we wouldn’t need fast write access
to random chunks of video memory; if we had fast non-sequential write
access to video memory, we could probably get away with so much of the
rendering happening in software (as OpenStep/Intel did). But with that
combination, I think we’re limited by the number of pixels we can push
through the hardware whenever we scroll or resize a window.
6. Despite Apple’s Switch campaign, all the recent survey’s I’ve seen agree that Apple has reduced its market share from 5% a few years ago, down to 2.3%-2.5% today. Is OmniGroup thinking of porting or rewriting their applications to other platforms? Is Windows or Linux on any
porting plans or you will stay faithful to the Mac platform?
Ken Case: All of our products are designed specifically for Mac OS X: we love
the combination of Aqua and Cocoa and UNIX, and we don’t have any plans
to port any of our software to any other platforms.
7. Do you think that Apple would make the big jump to AMD Opteron or to plain x86? Do you see something like this happening, does it make sense for Apple and maybe your business?
Ken Case: As you may know, NeXT supported four CPU architecture families:
Motorola 680x0s, Intel 80x86s, HP PA-RISCs, and Sun SPARCs. I’m sure
that Apple is keeping their options open, and will introduce support
for another CPU architecture when the time is right.
For this to happen, of course, they need to wean people away from (the
unportable) Mac OS 9. Since they’ve just introduced hardware which
requires Mac OS X, I think they’re one step closer to the point where
they can make such a move without backlash from the wider Macintosh
community.
8. OmniGroup also ports games to the Mac platform. How is OpenGL performance on the Mac today? How easy is to port 3D games from the PC to the Mac when the latest PC game releases have crazy hardware requirements even for PC standards, while most Mac users still run on Macs with slower graphics adapters and less overall speed than some brand new PCs? If this is not exactly an issue yet, do you believe that porting a “high end” 3D game from the PC in a year from now would be quite a challenge to optimize it in a way that it would run adequately to the Mac?
Ken Case: We have some great graphics cards on the Macintosh now, and potential
OpenGL performance has improved quite a bit with the release of Mac OS
X 10.2. But we’re still working under the handicap of relatively slow
CPUs and memory access, so until that situation changes it will always
be a challenge to make games perform well.
(As for running the latest games on older hardware: well, you can’t
really do that on the PC side either–in fact, that’s pretty much the
only reason I occasionally buy a new PC. So yes, it’s a problem, but
it’s not one which is unique to the Macintosh.)
9. How OmniGraffle 2.x and OmniOutliner 2.x have been selling? I read that in 2001 the company was not profitable. How was 2002? Easier or rougher?
Ken Case: We’ve made a pretty huge transition in our business over the last two years, from making 90% of our revenue from consulting to making 90% of our revenue from our commercial software products, and our boxed
software first appeared on shelves at CompUSA and The Apple Store within the last few months. We’re very happy with the success of our products (the last half of 2002 was definitely much easier than the first half), but it’s still a little early to judge how well they will
succeed (especially in the retail market), and we still have a lot of work ahead of us.
10. Is OmniGroup working on any brand new applications
for OSX or you are working on updating the existing ones?
Ken Case: Our first obligation is to our existing customers, so over the next few months our engineering effort is focused on major updates to our current titles–but once we’ve done that, we certainly do have some new applications planned!
I’m impressed with just how professional Case came across in this interview, especially in regard to the competition posed by Safari. I’ve always been a fan of Omni products, however, even though I’ve purchased OmniWeb, I use it rather infrequently. I’m looking forward to these updates he spoke of.
Any thoughts on how wise/difficult it would be for Omni to move over to KHTML?
Very nice interview. I think OmniWeb is one of the best browsers out there and I would really like to see it ported to Linux, I think it would do very well. Until then I guess konqueror will just have to do. To me tho it seemed OmniWeb was more stable on NeXTStep.
The OmniGroup is handling this situation with Safari very well, much better than can be said for, say, Opera. Despite the fact that I’ve always liked Opera, I find their whining about Safari intolerable.
Dohnert, I don’t think it would do well on Linux at all. To start with, OmniWeb focuses heavily on trying to use OS-supplied functionality to the fullest. A lot of the GUI-related functionality it relies on just isn’t there in Linux. Also, very few commercial products make it on Linux. It’s just a fact of life. People expect a free system to have free program. Sure there are some that make it, but they tend to be very powerful program doing a well-defined job. Another entry into the flooded browser market certainly doesn’t qualify.
It was great to read this – I’ve used NeXT and use OS X and just about all OmniGroup software. They really do make quality software. Very interesting and frank remarks about OmniWeb too. It really is a wonderful browser. Of course, I’ve been using Safari mostly since it came out, but still use OmniWeb. When they finally overcome its shortcomings, I would think it will end up being the most full featured Mac browser.
I’ve stuck with OmniWeb despite its obviously inferior rendering engine. One of the main features which has caused me to do so is banner ad blocking, a feature which is largely unique to OmniWeb. OmniWeb is one of the most customizable browsers I’ve used, and has the toolbar Opera wishes it had. I’ve also noticed that for web pages which don’t make large use of CSS, OmniWeb actually seems to do a better rendering job than KHTML. KHTML still seems like your traditional hackish open source project, implementing too many new features too fast without first ensuring that what has been implemented so far works flawlessly.
I very much dislike Safari. The UI is abominable. I say this not only because Apple used the brushed metal look for it (which can be disabled with a haxie) but also due to things like the weird progress indicator behind the address bar, which I find incredibly bizarre.
I do hope the OmniGroup will finish a new engine in-house for Omniweb 5.0, and not take the easy way out (KHTML).
>I very much dislike Safari.
I really like Safari. For me, it is the fastest browser on scrolling and resizing pages (which I do a lot) on my G4 Mac. OmniWeb, unfortunately, is one of the slowest in this regard.
>The UI is abominable.
It can be better, but its main case is this: simplicity. I like simplicity.
> due to things like the weird progress indicator behind the address bar, which I find incredibly bizarre.
I like that. In fact, it makes sense. I think it is one of the things you only like when you are getting used to with time.
> A lot of the GUI-related functionality it relies on just isn’t there in Linux.
Not entirely true. OmniWeb is based on OmniKit, which is available in source code. While some of it is pretty much tied to Cocoa, it should be possible to port it over to GNUstep with reasonable effort.
A very nice interview. Very refreshing to see honest and clear answers and a complete absence of corporate evasiveness.
Makes you want to go out and use their products.
Any thoughts on how wise/difficult it would be for Omni to move over to KHTML?
If KHTML (or some other external rendering engine) has the necessary oomph Omni desires, then it would be a wise move. The main appeal of OmniWeb is in my opinion not the rendering, but the creature comfort it provides (cookie management, bookmarks, download management, yadda yadda), and its good integration into the OS.
I think that OmniGraffle is one of the best pieces of software I have used (trial version).
On the other hand I was non plussed by OmniWeb…
I will buy OmniGraffle now that I have a Mac again. I use Linux for most other stuff.
I really do like Safari too. It could stand a little UI improvement, but I think what Case said about it is true (and what Eugenia said too). If one thinks of it as the bundled browser that is totally light and fast and uncomplicated, then a browser like OmniWeb can be the more full featured browser. But they really do, as Case said, have a lot of work to do yet. LOL, I used 2.7 on NeXT. It had what has to be the most unintuitive bookmark set up imaginable. In fact, that carried over to the first versions for OS X to an extent, but they’ve done a good job fixing that part of it.
OmniGraffle and OmniOutliner are great!
I like his attitude, because he likes to challenge and not whine like what Opera did. I like when the company admit and honest that their product isn’t doing very well, like what Ken admits that Safari is better than OmniWeb, but they are going to work harder and try get their browser better than Safari.
Thanks for the great interview, Eugenia!
I think OmniWeb will do well with adopting KHTML, it will make their browser more versatile. I do think that Omnigroup including Ken Case are handling Safari alot better than the Mozilla guys and the Opera guys in fact as he stated they are looking at KHTML to see if it offers more benefits then Gecko, Omni was always drop dead killer app on NeXT and rendering on NeXT was great. I dont know if its the new interface issues that affect rendering on Mac OS X but to me Safari is a great ” work in progress ” Omni under OS X is becoming much better and I look forward to seeing what is ahead. As for the commercial software issues brought up by Owen Anderson, I agree somewhat. Porting software commercially to Linux in the past was always a chore. Cheapos, people who always want something great for nothing, have long plagued the landscape of Linux. But I think as more professionals and desktop consumers switch I think the future of commercial software on Linux is very bright. I buy commercial linux software and I find it to be much better than most free alternatives. I will go free with Software that is very costly like thousands of dollars, but Star Office 6.0 is much better and font management is much better than its Open Office.org counterpart, If photoshop was to be ported I would be one of the first buyers as well with OmniWeb. If it was to become available for Linux, I would be the absolute first to buy it. Unlike some people I find functionality to be prescedent over cost.
Dohnert, you’re right. Some people would pay for it. But the likelyhood that enough people would adopt it to make it a worthwhile business venture is small to say the least. The Macintosh platform has a very strong tradition of shareware. Linux has a very strong tradition of free software. The two are hard to mix. Mac users are upset when free software doesn’t live up to the standards of shareware or commerical ware, while Linux users usually prefer free and open solutions over anything proprietary and closed.
Chris, so they should require all people who wanted to use it to install GNUStep? No, that’s not really feasible. All also remind you that GNUStep and Cocoa are not 100% compatible when it comes to GUI functions. They are transparent at a base level, and share some of their GUI functions, but they’re not identical. No, even porting it to GNUStep would take significant effort for minimal profit.
they should require all people who wanted to use it to install GNUStep?
And why not ? you’ll accept to install kde, gnome, but not a “small” package as gnustep ? (a du -h here of /use/GNUstep give me around 30 mo and it’s possible to remove stuffs)
No, even porting it to GNUStep would take significant effort for minimal profit.
Frankly, I’m not sure. Of course, Cocoa isn’t totally compatible with GNUstep, but it mainly consists of some widgets not available, or in early step (mainly NSDrawers, NSToolbars). GNUstep and Cocoa are not sharing “some of their GUI functions” but MOST of their gui functions.
At least you could say that porting OmniWeb on GNUstep should be far easier than porting OmniWeb to a totally different toolkit (eg KDE or GNOME)
I looked at the gallery and it seems quite nice..
For unix I have tried Dia but its less nice.
Does any of you know if there is any related
opensource projects?
Thanks
I have to agree with some of the previous posters who mentioned the benefits of supporting the Linux OS. I don’t know the OmniGroup’s views on this, however, they’re probably as devoted to their chosen platform as most other Mac vendors. I personally use Macs, Linux, and Windows.
Objectively speaking, I think that Linux will be unbeatable a few years in the future. Right now, it’s still difficult to install software, drivers, and set up firewalls and many other annoyances, but because of the generous licensing and both the massive corporate support it is gaining (IBM, etc.) as well as government mandates and funding in many countries, it is set to become the “next big thing”.
Now don’t flame me, I’m not an unbalanced zealot, I’m just being objective. You can’t beat FREE, and you can’t beat the extra benefit of access to source code if desired, and you can’t beat the masses of volunteers and paid developers who are pushing it forward at such an incredible rate. One day soon, it will overtake the commercial systems in every way. Heck, if RedHat hired Eugenia to fix their UI mistakes, RedHat would sell twice as many copies as they do right now! It’s just a fact that Linux will eventually win out over the other OSes, and Omni and other application vendors would be wise to position themselves to take advantage of this trend.
Please remember, I’m saying this not as a zealot (I use Windows, Linux, and the Mac), but as an accountant. I look at balance sheets, income statements, and charts of accounts every single day, and I can recognize growth potential when I see it. The Linux movement on the whole has so many assets and so few liabilities that I can’t see them not winning out in the end, unless the government outlaws open-source totally. And even then, the rest of the world will still migrate off Windows.
I like Omni. They make well-designed and thought-out beautiful apps. But as Apple invades their main markets, they’re going to squeeze them off the Mac. There will always be a place for them, but will it be enough to survive and prosper? Microsoft would squeeze them hard off Windows, and Mozilla’s current popularity would squeeze them off Linux. But they have the best chance on Linux, because there’s less leverage being applied against them by the corporation who owns the OS and thus has an unlevel playing field from which to attack Omni. In 5 years or less, it’s very likely that Linux is where the money will be. It’s just logical.
…that what they think is right, the best, the only answer. Did you people read the interview? Ken has been exclusively NeXT/OS X for over 13 years now! THIRTEEN YEARS!! He makes his living consulting and building high-quality, platform-tuned niche applications. This has been a working strategy for some time now.
You fools are asking him to abandon everything he has learned over THIRTEEN YEARS to produce giveaway-free software for a moving target of a platform exactly at the point where his company has developed recognition and is growing?
Some of you are just pathetic if you think you can warp reality with your delusions to the point where Omni is “better off” producing non-NeXT/OS X ports. Get over it.
…and, oh, while I’m at it, BeOS is dead too!
Omni could use the kHTML engine. Apple is currently trying to make the engine into a stand-alone Frame Work. Once this is done, any program can use the kHTML engine as a plug in. Thus Apple isn’t trying to kill Omni or Gecko; if they want to use the same engine as Safari, then they can. The real battle for users will be who addes the most extra features and has the best user experance.
Try using the IE engine in your programs under MSFT Windows. If you every find out how, MSFT will change the interface to keep you from doing it again.
I “switched” from WinXP to MacOS X a few months ago when I bought a new iBook. I was looking for a platform that reminded me of my experience with BeOS R4 but had modern hardware support and good software support also. This is not to say I have abandoned x86. I have a couple a WinXP machines at home and they do sterling work as Office workhorses and IMAP e-mail servers.
I have to say my experience of MacOS X so far has been excellent. My iBook interfaces with my home 802.11 network flawlessly and I can also plug it into my work network (British NHS hospital) and connect to the servers and internet proxy server. My only wish is that Microsoft would enable Entourage to work with Exchange Server – I await the rumoured “Entourage Enterprise” release.
Anyway that was a bit OT.
As part of my work I produce clinical guidelines for my hospital. As you can imagine this requires the production of flowcharts. I searched the net for sometime. There were, of course, lots of choices for x86, but very few for MacOS X. I then stumbled across OmniGraffle. I downloaded the trial version and obtained the free 24hr serial number. After this it became apparent that this application fitted my needs exactly. I have purchased a full license for the application and cannot commend it enough.
It is immensely powerful, yet easy to use. Its EPS and PDF export options allow me to import the files into InDesign and have the files produced professionally.
This application has allowed our hospital to treat our patients more efficiently. I hope Omni gain the success that their quality software would suggest.
No one is asking him to abandon the NeXT/OS X platform, all we are saying is that he could expand that range. OmniWeb as I stated is a drop dead killer App and if he was to start producing for these platforms he would do very well. OmniWeb was always my first browser choice on Mac OS X I do not like MSIE for Mac and I do not like Netscape/Mozilla and I really disliked Opera. But for business reasons I have to use Linux and it would be nice to see OmniWeb. We can all do wishful thinking there is nothing wrong with that.
There’s plenty wrong with it… It’s not productive if it’s a complete fantasy. And if it is, you sound like a moron who is not facing reality. And in this case, it is a complete fantasy.
“No one is asking him to abandon the NeXT/OS X platform, all we are saying is that he could expand that range.” And you are not hearing what I am saying. Omni is successful because it has a very focused strategy–focus on a platform with excellent OO APIs to produce applications that maximize that platform’s strengths. If Omni became a company that was OS but somehow tried to compete with proprietary apps like Explorer and Opera on all or some platforms, they would die. Have you read the comments? OmniGraffle gets more kudos than OmniWeb in many cases. They also have OmniOutliner and other apps. They are not just a browser developer. And what’s so killer about OmniWeb? I’m not saying it’s bad–I’m a user and it has unique features, but a killer app? Come on! Get with reality! Have you even used a Mac or Omni software?
This is what I am talking about… You are deluded, ignoring reality, and presuming that what you want, not only does everyone else wants, but would also be a winning strategy. That’s a freakin joke.
Omni should stay a OpenStep/OS X developer/porter/consultant. Plain and simple. You’re unrealistic and impractical “wishful thinking” will never change that reality.
Ken Case: I think that the big limitations here are that Mac OS X’s Quartz interface is currently software based (since no hardware exists which supports things like rendering nice antialiased text), and that the PowerPC chipset doesn’t have a direct equivalent to the Pentium’s write-combining mode for video memory regions.
I hope very soon, Apple will drop support for PPC architecture and use something useful and modern for a change.
I came into Mac OS X late in the game. In learning the Cocoa API’s and digging around for old NeXT documentation, I stumbled across some videos demo’ing EOF and they showed how easy it was to map an object to a database record or table. This effectively abstracts out the database specific tasks into normal object tasks. It was a rather beautiful solution. GNUStep is currently implementing EOF solutions but to bring these API calls into a Mac OS X application and make it work is very difficult as much of GNUStep is still unfinished.
I was greatly dis-appointed to learn that WebObjects used to include this technology under ObjC but dropped it for Java and JDBC which doesn’t do me a lick of good. I would be willing to buy an Enterprise Object API as long as it fully supported Oracle, Sybase, UDB/DB2, PostgreSQL, MySQL, and even the legacy Mac databases of FrontBase, etc.
Now I’ve got to build my own ObjC wrapper around PostgreSQL, etc. It’s just nowhere near as easy as it could be.
He was asked about porting and gave a somewhat indirect answer. Omni products are Cocoa based so they only run on NeXT, OS X, and with a bit of work GNUStep. They are based on frameworks that preclude them from running on a platform until the frameworks are ported to that platform.
Really, I think that Linux advocates should throw their weight behind the GNUStep project. This would allow Linux applications to be created easily and quickly (like Cocoa programs) and would allow OS X developers to port their apps to Linux easily. Although there is nothing inherently better about Cocoa, applications that use it are better because they require less effort to get it right. The development time is less than 1/10th the development time of other apps and that leads to better quality control, fewer bugs, and easier optimizations. Any developer will tell you that the shorter the code is the easier it is to write the app and fix problems.
I hope very soon, Apple will drop support for PPC architecture and use something useful and modern for a change.
Apple isn’t going to switch ISAs for a long, long time.
Why should anyone give it the time of day? GNUStep has abandoned PowerPC development–it’s original goal has been subjugated by uncompromising political ideologies… it holds no interest or value to Apple users despite the claim of bringing free software to Apple users. GNU/Darwin is a joke is far far as it is now just a crippled port of an experimental x86 Darwin port with little to zero hardware support. Since they refuse to compromise on their ideologies GNUStep cannot interact with anything interesting… it has no libraries of its own of any value…
By the way, what’s ambiguous, or indirect, about this: “All of our products are designed specifically for Mac OS X: we love the combination of Aqua and Cocoa and UNIX, and we don’t have any plans to port any of our software to any other platforms.”
That’s about as direct, concrete, and specific as it gets in this world.
Only the delusional would call that an indirect response.
Try using the IE engine in your programs under MSFT Windows. If you every find out how, MSFT will change the interface to keep you from doing it again.
I remember a mate showing me this new Windows web browser back in the late nineties. It was called something like Starbrowser or Startrekker or something like that. Something astronomic. Well, it was basically just a new interface to M$IE. Microsoft seems to have made its Internet Explorer API quite open, since that’s not the only example of reuse of its engine. WinAmp includes a “mini-browser” in its 400 KB minimal download. Do you think you could fit a complete up-to-date browser _and_ an MP3 player into that on the Windows platform? Nope, they’re just smacking M$IE into a WinAmp window.
So it would seem that M$ are more or less actively encouraging re-use of their browser engine. It boosts third-party productivity, if anything.
Thanks for providing us with insight into Ken Case’s view on OSX/Safari and the future of OmniWeb. Clearly the man is focussed , enjoys the challenge, and eager to find the perfect balance for both browsers to exist.
This interview should be emailed to the fine makers of Opera
“This would allow Linux applications to be created easily and quickly (like Cocoa programs) and would allow OS X developers to port their apps to Linux easily.”
How can OS X apps be easily ported to Linux via GNUStep? An OS X app should utilize CoreFoundation, CoreAudio, CoreVideo, WebCore (in the future), NetInfo DB, the OpenGL libraries if 3D is involved, OpenPlay, Rendezvous… Not a single one of these libraries will exist in GNUStep (well, WebCore maybe). Starting to see the problem? GNUStep cut itself off at the knees by being GPL Nazis (well, I can’t blame them–it is a GNU project after all, but that just goes to show that EVERYONE should have seen this.) GNUStep doesn’t provide a path to migrate Cocoa/OpenStep apps to Linux.
As for the other direction… How long is it going to take for GNUStep to reverse engineer all of the NeXT libraries to GPL code? It took NeXT/Apple 15 years; GNUStep is going to take a long time to catch up, and by then it will be a completely different monster than Cocoa…
In other words, GNUStep does nothing to create a bridge between the GPL and Apple; it is just another fork in the GPL world. That’s it.
I use Windows, Macs and Linux too and I can think of several other scenarios that could occur. I would very much like to see Linux come together in a formidable way, but I don’t see that happening. It is Linux that will have to radically change for that to happen.
—
In 5 years or less, it’s very likely that Linux is where the money will be. It’s just logical.
—
In 5 years of less, linux is where the money will be, but it will be a server. Not a desktop OS. IBM and SGI are not dumping millions of dollars into Linux to be a desktop OS. They are using it to leverage their low end servers. There is a big community push for the desktop – however so far that has been getting no where fast. I do agree with you that linux will go somewhere. But I don’t think it will be to a market that is anywhere useful to Omni who makes consumer applications.
I do have a general question, why is the point of “going with linux” from a business perspective, I do not see how IBM, SUN or SGI get anything out of GPLing any of their code or contributing to linux. Unless I’m missing something and they are not “supporting” the linux distro they sell you – it would seem like the same hassel as supporting their Unix.
Dohnert,
It would not expand his business, just explode his email inbox. There would be an unmanagable surge of email from stallmanites and linux-brats demanding he either release the source code under the GPL, make omni’s products cost nothing, or to make custom packages for X distro because Geek-Y uses it and would swear on his mother its the best distro in the world.
Hey, would you mind telling me what a modern platform is?
Do you mean the Itanium, which is so modern that nobody can really code for it well yet?
Or did you mean the super new non-crufy x86 architecture? If so we might as well just go with the ultra modern VAX designs eh?
Maybe the Opteron is what you are thinking? To put that one in perspective, Intel has been making x86 chips using IA-32 for how long? They invented how much of IA-32? If they didn’t go with a backwards compatible 64-bit instruction set, there must have been a reason. (No its not to be avante garde either.)
Do you mean the IBM 970? Which they’ve already said they are going to be using?
They port games. The ports are usually quite good, too, AND they keep up with them. MacPlay usually publishes them, I believe. This is likely very lucrative for them, because despite many recent improvements to the situation, there still aren’t quite as many commercial games on MacOS as are on Windows/DirectX (I would like to nominate the preceding as “Understatement of the Year”). Hell, there’s another company, Westlake Interactive that specializes solely in porting A-List games to MacOS.
Many, many people like to be able to buy A-List games for their platform of choice, regardless of what that platform may be. They could theoretically support their other software efforts (and consulting services) through whatever piece they get from game ports.
All they’d need to do is win a bid for something huge (like Doom III, Neverwinter Nights, Star Wars Galaxies Online, The Sims Online) once in a while and they’d be fine for a bit. Of course, actually doing so is difficult, as there is competition, but they have a nice history on their website.
Oh, and I think it would be very much nifty if they move over to some improved HTML parse and render engine by (at worst) MacOS 10.3 (or “Puma” according to Thinksecret). I don’t care if they license it, port or adapt another existing free engine, use KHTML/WebCore, or write a new one from scratch [difficult and unlikely, but possible]. I want better Omniweb, dangit, and I’m willing to pay for it.
–JM
Refurbished G4 towers are a GOOD thing, if you can find a reputable dealer.
TO: GNUStep is possibly the biggest joke in BOTH the Mac and Linux world
Why should anyone give it the time of day? GNUStep has abandoned PowerPC development–it’s original goal has been subjugated by uncompromising political ideologies…
Excuse me, but what the hell are you speaking about ??? note : I’m currently running GNUstep on debian/ppc FINE, thanks. Oh, you perhaps missed something… it’s GNUstep, not GNU/Darwin or whatever. Check http://www.gnustep.org , it’s a free implementation of OpenStep specifications.
TO: Sean, Don’t Buy Into GNUStep Idiocy
How can OS X apps be easily ported to Linux via GNUStep? An OS X app should utilize CoreFoundation, CoreAudio, CoreVideo, WebCore (in the future), NetInfo DB, the OpenGL libraries if 3D is involved, OpenPlay, Rendezvous…
An OS X app must use COCOA. As Cocoa is an evolution of OPENSTEP, and GNUstep being an OpenStep implementation, it’s largelly feasable to port Cocoa apps to GNUstep and vice-versa. Check http://www.collaboration-world.com/gnumail for example. A vast number of OS X app only uses COCOA — for the others, yes, GNUstep is not the new silver bullet, of course. But you could note that OpenGL is done via Cocoa (NSOpenGL), and it’s available also on GNUstep (in alpha state currently, ok). And that Rendezvous is OpenSource…
Not a single one of these libraries will exist in GNUStep (well, WebCore maybe). Starting to see the problem?
As I said earlier, it could be a problem, of course, but many apps only uses Cocoa…
GNUStep cut itself off at the knees by being GPL Nazis (well, I can’t blame them–it is a GNU project after all, but that just goes to show that EVERYONE should have seen this.)
Wow, smoking crack, anyone ??? is everybody confusing GNUstep with another projects or what ????? GNUstep is LGPL, thus you could link with whatever you want. Where do you see GPL “Nazis” (I should add, PLONK!) in the GNUstep project ???
GNUStep doesn’t provide a path to migrate Cocoa/OpenStep apps to Linux.
Yeah, right. That’s why I have already a few apps running BOTH on Mac OS X and on GNUstep right now, I must be dreaming, that’s it.
On a side note, a Windows port of GNUstep exists, in alpha stage, but it seems to works.
As for the other direction… How long is it going to take for GNUStep to reverse engineer all of the NeXT libraries to GPL code? It took NeXT/Apple 15 years; GNUStep is going to take a long time to catch up, and by then it will be a completely different monster than Cocoa…
Excuse me, but the biggest part of the work is already done.
GNUstep’s AppKit (-gui) is still a bit “young” but it works, period. And Apple/NeXT didn’t exactly reverse engineered the OpenStep framework during 15 years : Cocoa is basically OPENSTEP plus some occasional widgets additions (NSDr awers, NSToolbar..). The OpenStep specs were out in 1994.It was a job done by NeXT and Sun, and it was an amelioration of the previous NeXT framework.
In other words, GNUStep does nothing to create a bridge between the GPL and Apple; it is just another fork in the GPL world. That’s it.
You are definitely speaking of things you don’t understand, or you seriously misunderstood GNUstep. IT’S NOT GNU/DARWIN AT ALL, AND HAS NOTHING TO DO WITH GNU/DARWIN.
Please, Check