The Global ARCHICAD Community

Stay informed. Get help. Share your knowledge.

Discussions closely related to ARCHICAD. (Example: Do we need a Linux version of ARCHICAD?)

Moderators: Karl Ottenstein, LaszloNagy, ejrolon, Barry Kelly, gkmethy, mtron, Csaba Kézér

User avatar
By owen
#152433
Rob wrote:In regards to 64bit mac version.... well there is more than just recompiling the code to 64 bit.
I would like to remind all that Apple's sudden decision to axe Carbon one year ago was a bit shocker for all developers (even companies like Adobe had to postpone its suite). GS (and in fact most of developers) were using Carbon for back-compatibility reasons mainly.
Look i understand completely that re-writing AC for 64bit is no small task. I certainly was not expecting it for AC13. Wish yes, but I have been an AC user since the late 90's and so have grown used to GS timelines. 10 year old wishes, etc.

However the fact that they have managed to move AC13 to 64bit on Windows and not Mac is IMO just the result of a bad, bad gamble/miscalculation on Graphisofts behalf. I will refer to:

ArchiCAD Wiki TechNote: Carbon & Cocoa, 64bit ...
Q: How did Graphisoft decide on the appropriate Mac OS X environment for ArchiCAD?
A: When ArchiCAD was developed for Mac OS X we had to support a Mac OS 9 user base. In accordance with this ArchiCAD developers chose the only appropriate solution that is, the implementation of Carbon.

Q: What is Carbon?
A: Carbon is another application environment of Mac OS X. It was designed to provide a migration path for developers moving from Mac OS 9 to Mac OS X. If an application is written in this environment it will run both on Mac OS X, and on Mac OS 9

Q: What will be the situation after dropping the support for Mac OS 9?
A: Since Carbon runs programs just as well as Cocoa, there would not be any benefit in rewriting even parts of the code base in Cocoa. However, since Mac OS X supports compound program codes it is possible to use Cocoa for future development if it becomes more appropriate for actual projects.
So once Graphisoft decided to drop support for OS 9 in 2003 (in 8.1 .. ouch) there was no longer a requirement to use Carbon. It was purely a financial (i.e resourcing) decision to be made. So the gamble began during development for 8.1 ... 6+ years ago.
Q: Is Cocoa better than Carbon, since it is the 'native' application environment of Mac OS X?
A: No. Each provides different routes to the same core system routines. The choice of whether to develop a Cocoa-based or Carbon-based application depends on the purposes of the developer.

Q: What is the result of running ArchiCAD on a 64-bit processor-based system (such as Apple's new G5 (PowerPC_970) computers and Mac OS X 10.3)?
A: 64-bit processors offer the opportunity for programs to address significantly larger range of memory than the current 32-bit processors' 4 Gigabyte limit. ... Current computers do not exploit the 4 Gigabyte memory limit. Even so, Graphisoft keeps up with the changes and is closely following the developments of the 64-bit environments.
Bzzz. Turns out the correct answer to question 1 was 'Yes' because Cocoa was native and Carbon was the 'backward-compatibility mode'. Who would have thought Apple would eventually begin phasing out their legacy API's after 9 years. Maxon did, but not Adobe or Graphisoft.

And the second Q&A is just funny given the situation 5-6 years down the track. I mean if software development needs to be looking into the future then who would have thought 'well no machines TODAY have more than 4GB RAM anyway, so why would we need to support it'. Anyone with an eye on HW+SW development knew these exponential RAM increases were on the books and the major OS authors were already well on the way to supporting them.


Anyway, I just think this situation is purely the result of a bad gamble taken by Graphisoft(+Adobe) ... that they could keep on using a legacy API for years rather than bite the bullet and port their applications to OS X native API which would be guaranteed to be 64bit when the time came.

It was purely a money decision which is particularly galling given they are charging OS X users the same price for a lesser version of AC than they have developed for Windows.

Can someone explain WHY developing a 64bit Windows version was so much easier? Was it simply that there was no gamble to be made in API choice for developing on Windows?

If this uncertainty was the problem on OS X then Apple does have to take some of the blame, however as a software developer i would have thought one of the two choices would stand out as a sure bet, long term. They just took the easy way out for the short term.

Rob wrote:how many of you use more than 4GB memory for one project... I mean honestly.
Well at my last big office that memory error was not all that uncommon actually. Big models and people love having multiple copies of AC running with various Hotlink modules & TW drafts open. I know no model is even close to a 1GB file but there are all sorts of factors that add up to AC approaching its memory limit.

Even if the 4GB limit is not that common a problem today (overall users), can anyone honestly say that it will not come soon, and that AC would not benefit from handling as much RAM as you can throw at it?
User avatar
By PB
#152438
Owen, I agree entirely with your comments above. Mac users are, unfortunately, paying the price, in more ways than one, for GS having decided to concentrate on the PC platform. Whether or not this was the right decision is open to debate, although I concede that AC was probably unlikely to have survived without being written for PCs.

Within a medium & long term perspective, the progress made on TW is no doubt very important - assuming that it allows GS to consolidate & expand its user base amongst the larger firms.

What is galling though, is that I originally chose to invest in AC (version 6 if memory serves) because it was developed on & for the Mac (please correct me if my memory is playing tricks on me...): Now, development for the Mac has taken an evident back seat to the PC$.

The various other improvements included in AC13 are welcome, but side-step entirely the real needs of (almost) all AC users: A solid modern modelling engine, that does not, for example, require you to clean-up elevations/sections that have curved elements, SEOs that show correctly on plan, etc....

And, for good measure, the 4GB 32bit limit is actually now quite restrictive on a number of applications, including AC for many in the creative and/or technical fields.

Whilst I regrettably harbour the impression that my £600 subscription for this (very difficult) year has subsidised the large practices, my investment in AC (time & money) means that I hope & pray that GS make (& have made) the right decisions for the immediate future on the course of development for AC.

Perhaps most ironic is that, now the Mac is more popular than ever, developers who 'jumped ship' a few years ago are now being seemingly caught wrong footed :oops:

Oh well, at least this has saved me the angst of wishing for a nice shiny new 16GB 8 core MacPro for another year......

It seems to me (albeit slightly biased) that OSX is moving even further ahead of Windows/Vista/XP......I really, really hope that GS are going to take advantage of the technologies being made available in Snow Leopard to benefit AC.
Last edited by PB on Wed Sep 02, 2009 12:56 pm, edited 1 time in total.
By oreopoulos
#152439
Braza wrote:Well... I guess now "Small Fish" have no excuse to don't grow. :wink:

Again... It's a strategic release... not only for GS... but also for all AC users... Think about the opportunities on the horizon. 8)

It is a fair release... once "You always have to break the eggs to bake a cake", right? :wink:

And to some of you folks... lets not underestimate the huge effort behind TW2... its not fair. :roll:

a) I dont underestimate the effort. Its a big step. Its a great step. But its just for a portion of the users of AC. Not for every one. Only those that work in large firms and use it.

b) Brake eggs to make a cake? Correct. Charge extra for TW2 and charge for accessing the GS servers. I dont care. But take money from me... to finance the project that benefits big users? That is ridiculous.

c) Strategic decision. Yes . I get it. But its not free, and i have to pay for it
User avatar
By Rob
#152440
Well, I do not think that money reason is so small considering switching whole OS platform from PowerPC to Intel (and all developers had to switch their coding software too) then discontinuing the system library that BTW handles routines at the very low therefore fundamental levels. In other words your code would be ridden by those everywhere. And thirdly SnowLeopard came out on Friday so you expect everyone to be ready right that second. And all of it had happened in what 3 years or so!?... common mate... with the software of AC size.
Big models and people love having multiple copies of AC running with various Hotlink modules & TW drafts open.
4GB RAM is limit for ONE application instance - so if you had two ACs running at the same time both of them have separate 4GB addressing space managed by Memory Management Unit.

I am working on 140mil (500 units) project with really detailed model - barely touching 2.5GB in memory.
anyway I am not going to argue about technicalities but as I have mentioned before there could be a potentially interesting outcome if GS could manage to use any of SL's new technologies and I am happy they did not rush it.
User avatar
By owen
#152446
My point is once they discontinued support for OS 9 with AC8.1 then there was no reason to use continue to develop using Carbon other than financial ... 'we can't afford to resource the rewrite of old code'. I agree that would be a very big factor but they surely know Carbon wasn't going to last for ever and so the wise thing to do would be just move to Cocoa. They could have started that 6 years ago during development of 8.1 ... 4 years before Apple pulled the Carbon-rug out from under them. I think 6 years is a long time even in software development timelines.

And i know the 4GB limit is for per running AC. The fact is that people were still getting memory errors from AC despite having 6-8GB RAM.
Rob wrote:I am not going to argue about technicalities but as I have mentioned before there could be a potentially interesting outcome if GS could manage to use any of SL's new technologies and I am happy they did not rush it.
I'm not disputing anything technical, just the current situation is the result of a (poor IMO) strategic decision years back on what API to use. I completely agree that 10.6 could provide some very interesting outcomes for AC, but this Carbon v Cocoa episode has not shown Graphisoft to be particularly forward-looking when it comes to OS X technologies.
By Christiaan
#152452
owen wrote:but this Carbon v Cocoa episode has not shown Graphisoft to be particularly forward-looking when it comes to OS X technologies.
Well I don't know much about that but the fact that they don't even produce a Mac friendly interface independent of the Windows version certainly says first class Mac support isn't exactly a priority.
User avatar
By François Chatelain
#152454
Christiaan wrote: Well I don't know much about that but the fact that they don't even produce a Mac friendly interface independent of the Windows version certainly says first class Mac support isn't exactly a priority.
Funny no, when you think ArchiCAD originally was a Mac only program ;-)

François
User avatar
By Rakela Raul
#152487
another thing not mentioned is that MEP / ECO /+ will be upgraded for free..
:shock: :shock:
(hey ! im not a beta tester :lol: )
User avatar
By Djordje
#152491
Bricklyne Clarence wrote:There's a medical term for people who feel compelled to do things that they claim to otherwise repulse them. Do you know what it is?
You should know, as that's exactly what you are doing, complaining at no end about anything in ArchiCAD.

I am hereby publicly asking Graphisoft to include you, oreopulos, Necko and other whiners in the next beta test. Then you will learn what the beta testers do. You will see how much unpaid hours you spend working on the software that might die on you any moment, on real projects because that's the only way, preparing the reports so the bugs can be corrected, criticizing and discussing to great lengths the issues.

Would like to see you rant and insult people who did do that for you among others after that, whether you get something from the beta testing work or not. There are targets, you know. To be achieved in a certain time, so that crybabies like you will not ask why is your precious subscription that nobody forced you into about to expire and there is no upgrade, only to cry that the new upgrade is worthless. For EVERY upgrade (see oreopulos).

Get off your high horse, apologize to the people that you insulted and shut up finally. It is the almighty know-it-alls like you that brought this forum down.

Never judge a man before you walked a mile in his boots.
  • 1
  • 3
  • 4
  • 5
  • 6
  • 7
  • 12