Modeling
About Archicad's design tools, element connections, modeling concepts, etc.

!Restored: Musing on ArchiCad

KeesW
Advocate
Reading the stuff on Vectorworks on this forum is quite depressing. It does maybe 90% of ArchiCad for 30% of its cost. It will soon do 110% of AC and will (probably) still be much cheaper. We have in GS a company that is too complacent to attend major architectural and building conferences. ArchiCad still has crappy stairs, dodgy library objects and defaults that make me wonder if architects are still involved with its development. Users scream out for improvements in fundamental operations and instead get the ability to create multicoloured backgrounds and now, curtain walls.

I like ArchiCad and would like to be proud of it. But what, if anything, is it really good at? Is there anything at which it is best? All we get is compromises, clever ideas that nearly work but others that are never fixed. We have to pay extra for addons to do what our CAD system should enable to do out of the box.

Come on, Graphisoft, pull your finger out and really give us a first class product. Properly finance the fixing of long standing gripes before introducing yet another new feature that most users don't need. Listen to us, your users. Don't we have a say and don't we pay your wages?
Cornelis (Kees) Wegman

cornelis wegman architects
AC 5 - 26 Dell XPS 8940 Win 10 16GB 1TB SSD 2TB HD RTX 3070 GPU
Laptop: AC 24 - 26 Win 10 16GB 1TB SSD RTX 3070 GPU
18 REPLIES 18
Anonymous
Not applicable
All I can say is... wow. These guys have a bold development path they're undertaking. Can they do it??
Anonymous
Not applicable
I remember when I showed up at Studios Architecture and believed that Studios, then the largest Archicad user in the USA, would be doing BIM. Sorry. Everything they did was in 2-d. There was no BIM at Studios. What happened?

Just last night, I ran into someone from Pappageorge Haymes office the other night. They use Archicad... soley in 2-D. They do townhouses and highrises.

I used to own a license and found it to be really cumbersome when it was used as it was intended. So I use something else that is faster and simpler.

I'd like to see them really improve the software.

Make the 2-d and 3-d seamless. Make the stair tool work. Does it still have Plotmaker? It sucked.

And creating your own library parts like windows shouldn't be a chore.

Revit has taken over higher ed in the USA. You can find building parts in Sketchup and Revit. Where are the gdl models? the Archicad parts? There are few, if any.

I'd like to see Archicad succeed but Graphisoft has to step up and show us they can compete with the Autodesk juggernaut.
Anonymous
Not applicable
]One of my favorite charts in the keynote presentation was the one focused on describing where the competition was in terms of 3D modeling. (see image 08 above). This slide compared Vectorworks Architect, Revit, Bentley Microstation, and ArchiCAD, across a range of 3D capabilities, including in order:

Planer Surfaces: The basics of 3D modeling (think SketchUp)
Solid Modeling
NURBS - based modeling
3D Parametrics
If you look closely at the picture above you'll see that Vectorworks 2010, the next planned release, will have full support for all four types of modeling built into the application, and that with the exception of 3D parametrics, Vectorworks 2009 today has both NURBS and Solid Modeling.

That's impressive, but it's only PR exercise; look at the slide - it claims that ArchiCad doesn't have "3D PARAMETRICS" and they will have it in 2 years! Is it true? Well, ArchiCad has 3D parametrics - the problem is that the 3d parametric object creation is still at expert level. My problem is the cost of two last upgrades (11 & 12), which should be called 10.2 and 10.41, but the price is equivalent of full license of VectorWorks. MultiProcessor ability is great - but what we need is effective and easy "Visual GDL" and this will work for all sort of things, stairs, custom windows, etc... Models of type of interface are in now, a little bit of Objective, 3D Profiler GDL Tool Box and suddenly it may work.
......just to re-emphasize my recurring point on these threads as well as in the spirit of this particular thread; I will always continue to insist that Graphisoft absolutely need to completely overhaul their present and aging parametric-modeling engine which, as we all know, is a throwback to the days when you had to do half your input and changes by coding in commands and intentions through programming (gdl scripting) rather than via the traditionally intuitive graphical and architect-friendly visual interface/gui that most their competitors are now beginning to adopt. That's why it's so difficult to create custom objects from scratch, or modify existing library objects; That's why there exists a rather thin parts- and objects-library that hardly reflects the market offerings that most architects have to consider when spec' ing their designs. And that's why we're not likely to have any fully functional, or at least properly fixed Stair tool for a while any way.

And they are not likely to overhaul the engine anytime soon due to currently inflexible constraints that could have been avoided had they had a long-term plan with their clients needs first and foremost in mind rather than their shareholders' bottomline. The first is obviously their incredulous decision to switch from a 18-month to 2 year upgrade cycle to a yearly upgrade cycle for upgrading versions of ArchiCAD. A year is barely enough time to getting the program fully functioning with all the necessary hotfixes in place, as well as all the required updates and upgrades to complimentary third-party plugins and addons whose developers also seem to struggle with AC's new schedule. By the time most firms think it's safe to upgrade their systems to the latest version, a new versions is typically right around the corner. Shorter upgrade cycle means shorter Beta-testing phase, which inevitably means a more error and crash-prone product pushed out of the door. And don't even talk about the integrations of new features, which has now become an exercise in plugging in half-complete features from version to version and then completing the tool or the feature in the ensuing version. They should also allow us to pay half-license fees and complete the other half of the fees in the next version, no?

So it's no wonder they will never have the capacity to completely re-think and upgrade and overhaul the program's engine to adopt new technologies and enable more powerful functionality while they still struggle to maintain the current aging engine just working right from cycle to cycle.

Secondly, the lack of a long-term plan for the program as to where they want to take it in the industry, or even the lack of an articulation of such a plan if it exists, means they are not accountable for improving the software or anything beyond making sure the current version works well enough based on a (by now, rather low) standard they set for themselves going back to the version 8.0 debacle. As long as they can stay above that fiasco, then they consider any particular version a success. Such a methodology for running a development firm is fine if you're a front-runner (which ArchiCAD no longer is and is falling further away from), but when you're trying to be competitive and to give your clients an advantage, that's just a loony principle to adhere to. At some point in their history GS got stuck in some time-loop wherein they were the Industry's leading pioneer in this whole Virtual Building/BIM paradigm, and it seemed enough to just hold that mantle of being the Industry's first, and for a while, the only one as well. Well, it's no longer enough.

The point in all this is that, as elucidated in the 2 Vectorworks threads, they came to the realization that there's no going forward for them with their current methodology and modeling tools (despite the fact that VW modeling as it is, is still more robust than either ArchiCAD or Revit), so much so that they are adopting new technology to bring their clients and customers at par with AC and Revit users. So imagine the implications if you could if GS were to adopt some of the technologies shown in the video below of the Siemens parasolid modeling kernel in a sort of potential custom-parametric-object creation module to work within ArchiCAD. the video shows mechanical design parts for examples but the same technology and methodology has been used in Architectural contexts ranging from Gehry technologies Digital Projects, to Microstation and now even VW will apparently find a way to make it work for them. Some of those tools are just simply powerful and make complete sense from a design perspective. The ability to change the model and maintain updates using annotative functions such as the dimensioning and labeling tools, the ability to intelligently import foreign Geometry and natively manipulate it in your own work environment. Not to mention the ability to do all this parametrically in 3D with design intelligence built into the actual geometry as opposed to the objects. sadly, some of that is what VW users can look forward to by 2010, while we still hope for a fully functioning Stair tool.

the link:

http://www.plm.automation.siemens.com/en_us/campaigns/breakthrough/index.shtml


....the modeling video demos run from about a third of the way in after the intros by the VPs.
This was what I always thought that ArchiCAD would be capable of by the time the 21st Century rolled around.
Anonymous
Not applicable
Vectorworks may get parametrics and a change engine -- but are they going single file too? If they are, then this is a groundbreaking change. One of the limitations of VW is that their plug-in objects are limited in what they can create: the relatively easy workaround is that you can create your own symbol from scratch, but it's not parametric.

Revit is on a yearly cycle where certain new features are added during that development cycle; other features take much longer to develop and development on those features may continue through several years/cycles. So in any given release you're getting new features that were either developed entirely in that cycle, or were finished development in that cycle.

AC12 was a quantum leap for AC in my opinion -- something that should encourage AC users. A shorter release cycle doesn't mean that major changes to the software can't be made; it just means that Autodesk and Graphisoft can create some hype more often. And that's not a bad thing -- every March feels like Christmas to Revit users as they wait for a new release. Likewise, there are many Revit users who complain about the short cycles and how disruptive they are: so there are many who do not use the new version until 6,9 or even 12 months after the release has been available. Most Reviteers like to be bleeding edge though

I was talking on the phone yesterday with a AC-Talk member who's become a good friend thru this forum, and we were commenting on the fact that there is something wrong -- even dysfunctional -- with how ArchiCAD's development and marketing is being managed. Thanks to the internet and user forums like this, a potential software buyer is going to be able to cut through the marketing baloney that any given software developer puts out. An educated buyer is going to be able to gauge the technical competency of the software fairly readily, so it behooves the developer to make software do what potential customers expect, based on their needs and competing products' capabilities. For AC to stay on track, there needs to be some fresh blood in AC's management. Put THAT on your wishlist 😉
metanoia wrote:
............ A shorter release cycle doesn't mean that major changes to the software can't be made; it just means that Autodesk and Graphisoft can create some hype more often. And that's not a bad thing -- every March feels like Christmas to Revit users as they wait for a new release. .......
.......are you seriously drawing a favourable comparison between Graphisoft's switch to a shorter release cycle with Autodesk's yearly cycle? For that matter are you seriously suggesting that Graphisoft's has the necessary or even comparable resources to sustain a yearly release cycle to the degree that Autodesk have been able to do? Because, even with the Nemetschek buyout, it still isn't even close.

A shorter or yearly release cycle certainly works for Autodesk who can afford to throw almost limitless resources and however many coders and programmers they may need to make sure any given feature or tool makes it into the next release or within the 12 month cycle - especially with Revit being the software that they are marketing hyper-aggressively at the moment.
Graphisoft, not so much; simply put, they are stretched far too thin. And it shows with every new release that comes wit some half-baked feature or half completed 'big' new tool, which has to take another version or 2 to complete.

Every March or every new release may feel like Christmas for you guys , but for us its beginning to feel like "I wonder which, if any, of my wishlist features from 10 years ago won the lottery and got fixed this time around?", or "I wonder what new interesting but ultimately not-very-useful-for-my-day-to-day-work tool got integrated and which we'll have to hear them praising for the next 12 months as a paradigm shift in design?"

A shorter cycle only works if you can afford to compete completely in that time scale. Otherwise it just ends up hurting you in the longer run more than helping anything.
Anonymous
Not applicable
Otherwise it just ends up hurting you in the longer run more than helping anything.
So AC12 has been out for probably at least a quarter of its lifespan and we are still missing critical plugins - has the yearly upgrade proved to be nothing more than marketing?
I've posted before - why there isn't a one stop shop listing all plugins, add-ons etc and their status....

Apple have showed the way with their app store. A veritable one-stop supa-store. (realistically with GS track record if i made a tool I'm not sure I'd want GS selling it!)...
Anonymous
Not applicable
Bricklyne wrote:
A shorter or yearly release cycle certainly works for Autodesk who can afford to throw almost limitless resources and however many coders and programmers they may need to make sure any given feature or tool makes it into the next release or within the 12 month cycle - especially with Revit being the software that they are marketing hyper-aggressively at the moment.
I'm sure you're right on the resources front, but the principle remains -- you might get fewer features more often. We got the new renderer this cycle, which was a big job that took several cycles. 2008 was a much less exciting release. Rolling out Revit releases from cycle to cycle is very easy, though not so much with this release since you have to remap all materials from upgraded projects to a 2009 render material. I can see short cycles being bad *if* they really do hamper the developers and/or the new release is laborious to roll out. I suppose it is a pain for any significant large design company to have to roll out new software every year though...
Anonymous
Not applicable
metanoia wrote:
I suppose it is a pain for any significant large design company to have to roll out new software every year though...
This is why some firms prefer to skip a version and upgrade every other year. The transition costs can be considerable. This was very true for upgrading from AC9 to any later version (due to the PlotMaker integration) but has not been as much of a problem since. Most people I know have moved from 11 to 12 with hardly a hiccup.

Of course the annual cycle also tends to drive and be driven by the subscription approach Autodesk takes.