One of the announcements at the current OO.oCon is the decision to build a Cocoa-based version of OpenOffice.org 2.x for Mac OS X. This is a step forward compared to the current X11-based port and the Neo/J Office which is not up-to-date. No timeframe estimation was given for the completion of the project.
Native Cocoa OpenOffice.org Build Planned
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
its about time =)
amen to that!
If the X version isn’t up to date, why should a Cocoa version be? Won’t Cocoa have more code differences to the base version, which I presume to be Linux based?
Yes, a cocoa version would have a significantly different code base from the other versions. That is why carbon exists.
No a Cocoa version won’t have a significant difference. It will have Cocoa interfaces and depending on what the hell they do with advanced Cocoa APIs then they will have a varying codebase difference.
This won’t be a just a port of Cocoa wrappers around OpenOffice. Who cares? Cocoa has many advantages that can be leveraged or not to make OS X version actually better and less bloated.
The X version is up to date, AFAIK, but the NeoOffice/J (Java) version is behind. NeoOffice/J isn’t really that nice (slow) and using X11 just sucks, IMHO.
I’m just wondering how many years we’ll have to wait in order to see a working Cocoa version of OO.o. Maybe they should have done a proper Qt port so that both OS X and KDE could get a nice interface.
The official X11 version is at OOo 1.1.2.
NeoOffice is at 1.1.4 and does not depend on X11.
I really don’t want to purchase MS Office X for the little bit of word processing I do on my Mac, and I’ve tried AbiWord (getting closer, but still not all there), NeoOffice/J (this is the one I use on a day-to-day basis. Slow as molassis, but it works) and ThinkFree Office (not even close). Let’s not even talk about AppleWorks shall we?
I think a native version of OOo for the Mac would be excellent. Thanks!
ahem.. iWork is fantastic.
Well it is OK, for school work and the like, but iWork is more like a MS. Publisher replacement. Office, and Open Office, has other tools as well such as a spread sheet (You know the bit of software that everyone in business, but they shouldn’t). Also we need a good Free at least as in Beer word processor.
iWork runs like shit on my 1.5 GHZ G4 PowerBook with 1.25 GB Ram…. maybe even worse than OO.o …
I am excited about this, although I hope they keep it simple!
Kind of like the Gimp conversion SeaShore
http://seashore.sf.net
I have a 12in powerbook, 1.5Ghz, with 1.25 GB of RAM, and iWork runs great on it? I like Pages a lot. It’s actually a pretty nice combination of word/publisher.
It does run ok even on my 12in iBook, 1.2Ghz. However I have do admit I have a bit let down with the performance.
Knowing Apple had hired the people from Gobe Productive which was amazingly fast on much older machines …
On GoBe Productive (on BeOS 4.x) on a dual PII 350 mhz machine you could move elements in the middle of the page and the text would flow around it _very_ smoothly. Now it’s a bit laggy on a system that supposed to be much faster.
It doesn’t really make it unusable it just kills that ‘wow’ effect. When I first saw the demo in a keynote it was as smooth as Gobe and I expected it would be the same on my low end machine. It wasn’t the case.
I need to add that I was talking about Pages only.
Keynote is just awesome to get a great presentation done. Nice effects, all very smooth. Just great. (Though the flash export plugin could be a bit better).
Wow…
I have a Mac mini 1.42Ghz and 1Gb Ram, Tiger, and iWork is realy not slow, faster than Word in Office 2004.
I think it’s time for you to FORMAT and Re-Install your Mac…
You might give Mariner Write a try…it’s not free, but not expensive and seems to do a good job.
This is NOT, unexpected. Don’t you remember that the developers that worked at this said that they would wait for after OO.o 2 until they would continue working on the MacOSX native version? Most people complained that it would be so long before they finally got what they wanted. The developers didn’t wait till after OO.o 2, nice from them.
To bad most people complained back then.
It’s pathetic isn’t it? Cause they can run it just fine, it’s just not “pretty.” It’s pretty stinkin easy, you just install it with fink after installing Apple’s X11 (the xfree one is terrible, it doesn’t integrate at all). And you should have both of those anyway (I don’t know how anyone can live without x11).
I wonder how some developers do it. First they deliver the work, then the users cream, THAT SUCKS ITS UGLY. So the developer works on a better looking version, but (in this case) the underlying technology changes. So the developers say, we will wait until the next release so we wont have to do the work twice. To which the user say, YOU SUCK, I WANT THAT NOW. The developer dose just as he said and somehow the users still aren’t happy, you know what, DAMN USERS.
It’s not only ugly.
a) Neo-Office is DOG slow on a 1.2 ghz machine. It makes me want to buy a windows machine. Try opening a 200 pages document (like a book) and it’s really unusable. Like 50x slower than MS-word when repaginating.
b) OpenOffice 2.0 is *much* faster on the same machine. Yet running in X11 brings some real usability problems:
The menus are not in the right place (for OSX), the shortcuts are non-standard (for OSX), you have to use other fonts than you get in other apps and so on.
Windows users don’t have to suffer those problems, why should mac users?
On the whole the X11 version just feels bad. I can use it, and I do use it but a native OSX version will really make it good. Now I guess if the X11 implementation itself was better integrated (like it is in KDE or Gnome) you wouldn’t need a native version that badly.
>Windows users don’t have to suffer those problems, why should mac users?
Becouse you don’t have the developers. Do you know how long these guys been asking for help without reciving any? They didn’t get any thanks for their work, didn’t get any money, just abuse. Thats what suckes about it. If you can’t help them and don’t want to encourage them fror the work they did, atleast don’t abuse them.
If you don’t want to be abused, don’t write shitty programs.
spoken like somebody who doesn’t use a mac on a regular basis.
Native mac applications have a certain amount of functionality and system integration that users expect to have available to them. Things like drag and drop support along with .Mac, iLife (this is one reason why iWork is particularily nice), address book, etc. integration are not currently available in the X11 version of OO.o. Also, native quartz rendering would make a dramatic impovement in the visual appearance.
While it’s nice to have an app following the interface guidelines, it’s not just a pretty interface that mac users care about.
Have a 1Ghz G4, 1Gb internal mem and both Pages and Keynote run great. And I write manuals for a living, with a whole lot of screenshots, tables and such … maybe it is the ram.
I have tried to use OpenOffice, but on Mac it simply didn’t run very good. NeoOffice crashed my machine once. This Cocoa version is extremely great news.
> NeoOffice crashed my machine once
Just once? Geez, most programs I use crash a lot more frequently than that!. NeoOffice/J is slow, but quite useable on my Cube (450MHz G4, 640M ram). If you want a free word processor for the Mac, then you can hardly complain about this product.
Definitely get OO 2.0 and it will run much faster than Neo-Office. I can hardly believe you when you say it’s quite useable on your machine. You probably work on 1-4 pages documents.
1-4 page document? Yeah right… just finished a short manual counting 47 pages, including 311 illustrations, including 51 screenshots, most of ‘m including arrows and lines point to text. In all 6713 words, 1799 lines, 620 paragraphs. Totalling 43.697 characters.
Do know what you mean though, but that has something to do with tables or so. If you put in a table (even if you have a one pager) and you try to type some text above the table, Pages is slow as hell.
Didn’t know OO2.0 was available for Mac yet. Will look it up right away. NeoOffice was not my thing and 1.1.4 … eh … I can’t get used to that interface (my problem, I know).
“> NeoOffice crashed my machine once
Just once? Geez, most programs I use crash a lot more frequently than that!. NeoOffice/J is slow, but quite useable on my Cube (450MHz G4, 640M ram). If you want a free word processor for the Mac, then you can hardly complain about this product.”
Sorry 220.245.132.—
I’ll give more details: heavy kernel panic which required a reboot and my OS never booted again. Had to do a reinstall of that machine. Not some kind of crash in which an applications hangs or panics. No, they whole system went flat. Thank you, not my cup of tea.
>I’ll give more details: heavy kernel panic which
>required a reboot and my OS never booted again. Had to
>do a reinstall of that machine. Not some kind of crash
>in which an applications hangs or panics. No, they whole
> system went flat. Thank you, not my cup of tea.
Sounds like a kernel or driver bug. Don’t blame applications for OS level crashes.
Yep, a userland application shouldn’t be able to crash the kernel like that, so your problem was caused by Apple, not OO.o.
“Yep, a userland application shouldn’t be able to crash the kernel like that, so your problem was caused by Apple, not OO.o.”
Interesting, but that’s not my experience.
1. The totally updated production machine had worked normally for months, with only five (all registered) applications installed, besides the OS: Adobe Photoshop, MS Office, iWorks and Omnigraffle.
2. I install Neo/J and I get a kernel panic.
3. After reboot the machine panics again in the white screen (with the Apple).
So to me it sounds like:
1. Machine is perfectly okay.
2. I install some application (NEO).
3. The machine crashes.
1+2+3=the application crashes the machine.
Neo/J “caused” the crash, but the crash itself was due to a bug in the OS that Neo/J just happened to expose.
Yes, the bottom line is that Neo caused the crash, and QA testing should have caught that and added a workaround for whatever bug they’re exposing. However, the actual bug is in Apple’s code and I hate to complain about Neo when what their developers did “should” have worked.
Anyway, you might try again when a new version of OSX comes out, or when the next version of Neo is released. Of course, if you’re happy with the alternatives, then you can just use those.
I thought the open office mac team was just 2 men.
Can they pull off something like this without more manpower?
This is just hype.
They were working for ages on an earlier port to an OSX UI. It never happened.
This won’t either.
I am not a programmer so I am curious why a native port hasn’t been attempted sooner. Openoffice is open source and Xcode ships with Tiger. Why can’t you stick the source in xcode and have it crank out an Openoffice.app? Are there additional customizations that have to be hand worked before that can be done?
A native version of OO.o is not a recompile away. First you actually get it to compile with something (X11 in this case). To do that you sort out the non graphical parts, the build system and what not. Its not like Operating Systems are standardized on what they provide for programmers even with MacOSX being a Unix. Once that’s done OO.o runs on MacOSX so long you have X11, this is an important step, but now the hard part comes. The GUI you see needs to be talked to and the API it speaks is deferent from OS to OS. OO.o has an abstraction layer for this now what the programmer has to do is basically write the equivalency of all of that code NEW for MacOSX (or any other native port), that’s a hell of a lot of work. Why didn’t the programmers do that work before? First, there was only 2 of them back then (don’t know about that now), second, with OO.o the abstraction layer changed, so if they would have continued their work back then they would have had to do the work twice!
You can see there is a hell of a lot of work involved in getting OO.o working and Xcode doesn’t do any of the work, its all done by the programmer.
Same problem as when porting most programs. While the core code is rather easily portable all the GUI things are tied to the OS itself. Drawing a button on OSX is just not the same thing as drawing on Windows, same goes for every UI element and the way they interact.
Basically you have to rewrite (or write on top on some abstraction layer) all the UI elements. That’s a lot of work. I assume they have some abstraction layer from the start but it’s still can’t be done in a week.
You’ll find lots of developers for Windows because of the mass, quite a few on linux (because it’s Linux, everyone is a dev) but much fewer on OSX.
BlackJack75 and Anonymous,
Thanks for the explanation! Wish I had the talent. its sounds challenging, indeed.
Greg
Start learning python and pyobj, you will get the smallest fealing of what work is behind it but with more fun.
> I am not a programmer […]
I see.
> Openoffice is open source and Xcode ships with Tiger.
This is a frequent misunderstanding: “It’s open source and source code is machine-independent, so everyone with a compiler can compile the source code into a binary and that’s it.”
Jesus, please learn the basics. Source code is only portable if it uses standard libraries like the standard C library or the standard C++ library only, or if it’s an interpreted language and only build-in features from the interpreter are used.
Just recompiling the source code is absolutely impossible for GUI applications because there is no standard GUI! Every platform implements the GUI in a completely different way, therefore all GUI applications need to be rewritten for the target platform.
There are several ways out of this mess:
1. Rewrite your application for every target platform.
2. Use a library that wraps around the target platforms’ GUI and hides it behind a platform-neutral one, like wxWidgets.
3. Use a library that implements a GUI completely independent from the platforms’ natives ones, like Qt.
4. Use an API of another platform you already support on the target platform, even if it’s not native to the target platform.
What the OpenOffice.org guys did until now is variant 4. It works, but the users don’t accept it. The best solution is variant 3, but it’s not possible because Qt’s licensing terms don’t allow it. Variant 2 implies that the Windows and UNIX targets would have to be ported as well, which is too dangerous if we consider that the effort can actually fail. So it will be variant 1, this is the best from the users’ point of view, but the worst from the developers’ because OpenOffice.org’s GUI actually needs to be rewritten.
> I am curious why a native port hasn’t been
> attempted sooner.
– Because Sun is not very much interested in the Macintosh platform
– Because Macintosh developers are not very much interested in OpenOffice.org
– Because it’s difficult, time-consuming and expensive, given the fact that open source development DOES cost money unless it’s done by volunteers, which is not the case here
Good news, now I won’t have to use Office.
In the “article” the team talks about how it would be nice to get help from Apple. I really hope someone at Cupertino is listening.
Unless they are willing to do an ODF compatible word processor themselves (Pages 2.0 or even 3.0) they can only gain from it. And even if they want do do a similar Office suite it would still be good for the platform to have a decent OO version.
Right now if you try OpenOffice on Windows or Linux (both well integrated) and switch to a mac, you’re likely to come to this conclusion : “hey the mac sucks, the same apps feels bad”. Well, either feel bad with OO 2.0 and X11 or feel slow and one version behind with Neo-Office.
Apple won’t help. They have their own stuff (the rests of Claris Work and some new stuff) and supposedly have a deal with Microsoft about not helping OO.o but that is probably just paranoia.
You took the words right out of my mouth. If Apple helps OO.o, it does that at the risk of helping a product that competes with Microsoft’s offering. One thing that would be bad for Apple sales is to lose Microsoft Office.
Too bad Apple Inc. has given no support to this, it would be a lot faster in getting here.
Isn’t that the amount they had in spare cash now after a few years of successful iPods?
I wish they would just spend 100 milions helping improve some open-source projects (outside darwin…) so they will be better on the mac. Or offer some developers (equals to cash anyway) that would work just to make them more mac like and more performant.
Coming from a Linux/Windows machine I was a bit let down seeing that Eclipse on the mac was slower for example. It improved now, but it’s not because of Apple.
>>I wish they would just spend 100 milions helping improve some open-source projects…
Me too, how about improving GNUStep and making it and Cocoa AND Objective-C viable, crossplatform technologies? Look at Java with all its faults its still damn sucksesfull!
I’ve heard this before, and I’ll believe it when I see it.
From what I understand, a Cocoa port is a project of unbelievable complexity. OpenOffice.org is just too big, and its API is too different from Cocoa, to make this a realizable goal–unless Sun is going to start throwing as many programmer-hours at the Mac version as they do the Windows version.
This article (http://www.macdevcenter.com/pub/a/mac/2005/08/09/neooffice.html) is an interview with the lead (only) developer of the NeoOffice-J port. It’s a pretty interesting picture of the difficulty involved in just getting NeoOffice to run reasonably well on the Mac (mostly-native look-and-feel, no X, etc.). Patrick Luby uses Java to connect the GUI bits to native Mac API’s–he says it’s about 1% of the millions of lines of code. It’s been at least two years of work just to get it this far.
He may have gotten it as far as he can. The transition to MacTel is going to complicate things a lot. Java that hooks into native API’s (via JNI libraries) won’t port easily, and this affects NeoOffice. So does how Apple is implementing Java: the version currently used by NeoOffice (1.3) binds to the Carbon API, which is closer to OpenOffice.org’s API. More recent versions (1.4 and 1.5) are bound to Cocoa.
Though it runs very slowly on my machine, I use NeoOfficeJ for a lot of stuff, and I appreciate what Patrick Luby has accomplished. My guess, though, is that the current version of NeoOffice is going to be as close as we get to a fully native OO.org port for the foreseeable future.
This is the first time I’ve heard the desire and the will “to climb the mountain” announced, instead of being a cool afterthought “if someone would do it”.
My needs are modest to intermediate so both NeoOfficeJ and now the Beta 2 of OO for X11 serve me well. I find the interface relatively easy to figure out. Different, but relatively straight forward.
There is a lot of concern with “deals” with $MS. But Apple made a public committment to Java a few years back.
And I just hope they can adopt OpenStep’s text system and Apple to improve Cocoa speed. Webcore didn’t use the system coz Cocoa’s text system is slow and that’s why many useful Cocoa text features stop functioning in Safari (eg. word segmentation). I don’t see why Cocoa’s text system is that too slow, it is quite fast and very usable with GNUstep. This raises to me another question why everything in Cocoa (beside the eye-candy) seems to be slower than GNUstep. Someone has told me it’s because the CoreFoundation hmmm…
Study this document:
http://developer.apple.com/documentation/Cocoa/Conceptual/TextArchi…
Pay close attention to the Carbon Application Services interaction insteady of Cocoa<->Cocoa Application Services.
Once they dump Carbon the speed up will arrive.
Heh… here’s a quick way to “highlight” the lines of code which need to be changed…
Take the Windows code, then the X11 code… then do a Diff. Anything that doesn’t match exactly, rewrite for Cocoa.
No?
To much work?
Well, it was a thought. Sm:)e.
The GUI and backend should have been separated from the start. For an application that touts “portability.”
As said, about time!
Hooray!!!
Waiting patiently!
Just about 5 minutes ago. I saw NeoOffice/c on NeoOffice website.
By the way NeoOffice/J takes forever to load (about 10 minutes) on a PowerMac G5 single 1.8GHz 1.2Gbyte RAM machine, because of the FONTS! Have quite a huge collection of fonts in there. Need to remove most of those which I do not use.
http://trinity.neooffice.org/modules.php?name=Forums&file=viewtopic…
This tells me that whoever is doing this will have a long, hard slog. Perhaps the situation is better now that 2.0 is in advanced development, but it probably won’t be much easier.
This is super awesome news. I can’t wait for its release.
However, since the speed of open source development resembels evolution, I expect to see this hit alpha within a year or so.
I don’t know what you guys do, but if I am multitasking and need a word processor, Pages just crawls. Word is *so* much faster.
And I doubt its “bad ram” because nothing else bombs my machine so badly like Pages did when I tried it. That was when I had 768 MB ram, now I have 1.25, haven’t bothered trying it again because if its going to run that horribly its not worth using IMO.
Abiword is nice, but its interface is just very cluttered…. its too much like a Linux or Windows app.
The current official version of OOo for OSX is at version number 1.1.2
NeoOffice has released a OOo 1.1.4 based version a while ago and has just updated its code to Java 1.4.2.
Now that OOo has a 1.1.5 version NeoOffice will certainly move to this new code base as soon as it is released for OSX.