"New" Global Libraries - Parametric Object Creation?

[Kalib Stewart]

There's been a lot made about moving to global libraries, which hasn't actually done much, seeing as the Archicad libraries have barely evolved over the past decade... Still no claddings, still no parametric windows and doors, still only like 2-3 skylights, still very very limited fixtures, casework, and appliances. Half of my Toolkit is just CI tools and their Select Library.

With all this fuss about transitioning from one form of library pack to another form of library pack, are there any plans to finally allow us to create our own "parametric families (objects)" in both 2D and 3D? 

GDL is archaic and redundant for anyone trying to be productive design and documentation-wise; we don't have the time nor the desire to code so that I can make a 3D tap or a stretchable 2D symbol...And then test it for hours on end until I figure out where the single wrong line of code is.

Param-o is clunky and requires creating 10 blocks with 20 lines to create a four-leg table... Nice if you're experienced in Grasshopper, worthless if you're not.

And, the library part maker, at least in my experience, seems awfully slow and limited. There's also little to no available documentation available on how to even use it in an optimal fashion, and it seems I have to be doing it in separate files all the time.

So, for this 28 release/preview, are actually going to see any substantial improvements in our libraries or our library creation ability upon the release or before the end of this decade? This global library "feature" has so far been nothing but moving existing objects into new folders, and I could have done that in Windows Explorer and some website downloads.

This "Global International Library" has given me access to maybe 4 new useful objects...

It's been what, 10 years and for example there's still only one very limited vanity option.... 10 years with essentially the same limited library and no efficient ways to add to it. Is there actually any work at all going on in this area at GS?

Revit's had this mastered for years and it seems GS hasn't even started.

[claudiuiancic]

I would like to see improvements also in attributes management.

[stefan]

The flexibility and power of GDL is far beyond what you can achieve in the Revit family editor, which has like five main modelling tools. But the barrier to entrance is high for GDL. Coding objects is not for everybody and graphically sketching a basic Revit family is approachable to more people (but also typically a task for an expert BIM modeller). But when you look at any reasonably complex GDL object (those which have multiple pages of configuration - with graphical icons and adapting UI, not a flat table of parameters), there is no way this could be set up in Revit, not as efficient, not as a small object file, not with a graphical interface, not with interactive hotspots and preview generation etc... 

Library Part Maker, PARAM-O, save object to GDL etc... are all approaches to make it a little easier. But the real power of parametric objects can only be reached through the coding of their behaviour and internal logic.

You could argue that Grasshopper is one of the best visual programming environments, but the more I use it, the more I start adding Python nodes to cover stuff that would take too many separate nodes in an ever expanding network.

There is no "one size fits all" solution here. Some people are fine with a static object that they use once (model with slabs, walls, shells, morph and store as object). Others need some elementary parametric behaviour (PARAM-O seems to cater to this, but I agree you need too many nodes for even the most basic table). And experts may learn GDL and work from there.

One final limitation: trying to "improve" upon an existing object or expand its behaviour (e.g. to have cabinets which fit under a sloped roof or a non-rectangular corner) is al but impossible for the existing objects in the library. They simply have too many complex functions and macros, which are only properly understood by their maker.

[abdelaziz]

Hello,

I think there is still a track to experiment on the side of the philosophy of Cinema 4d and its module "MOGRAPH"

which consists in partitioning the Functions (set of codes executing a general function) into 3 categories "parametric or static element" Object, "Generator" and "Modifier", and these functions must always be placed in a "Parent and Child" system like folders and subfolders, the generator must always be placed in parent of element, the modifier must always be placed in child of the element or generator which allows to act on the static element to make it parametric

[Kalib Stewart]

Thank you. I understand all of this, but 99% of the time we don't need that level of complexity. We need efficiency and simplicity. Sometimes I just need to model something bespoke that I regularly use, but I need it to be flexible and adaptable too. Widths, heights, thicknesses, symbols, types, materials etc; 

We spend enough time designing, drafting, coordinating, learning multiple softwares, CPD, and navigating regulations etc. Learning GDL is simply not worth the additional time, nor is it practicable for small-medium sized firms. On top of that, it is slow and clunky, oh so slow.

The Revit family creator is approachable, easy, and efficient for those of us working in the 3D/2D environment. If I want to create new furniture, windows, symbols, and detail items I just do it on my second screen (another gripe with AC) with the same architectural tools and mindset while also working in my project. I draft it in 3D and plan/section/elevation, set my parameters, formulas, and constraints and place it in my project. Then if need be I can also set it to my project's global parameters.

If it doesn't work I open the "object" file, and tweak it as needed. I don't scroll through 50 lines of code wondering why my plan-view text is in the wrong place, or why I can't adjust my leg thickness to scale appropriately with their heights.

It is beyond absurd we still don't have an equivalent simple and easy method for both 2D and 3D archicad object creation.

As architects we work with drawings and models, there is absolutely no need to add the requirement to learn code into the process; and until GS addresses this, manufacturers, suppliers, cabinet makers etc will continue to only provide rvt files to us and architects and engineers will continue to gravitate towards Revit for bespoke BIM modelling. This has been an issue for the past decade and there is no excuse as to why we don't have the ability yet.

I haven't even started on the mesh tool yet either...

[Jman]

There is 100 and 1 ways of doing objects in Archicad some of them simple like Revit family,

GDL scripted object is 99% perfect and can have over 2000 parameters. It can be animated in Archicad and represent many version in 3D and 2D. 

Revit families don't have those options, Last time i did work on Revit families were a bit dumb in comparison. 

I guess you did spend time learning how to make families, so if you want to climb next level of Archicad experience without "limits" than learn a bit of GDL When you make one, you might like it. 

I hope Graphisoft will improve script editor but not replace it with anything Revit like

[NigelMX]

I agree with Kalib, we shouldn't have to learn all this other stuff to be able to design what we want. even if we just look at windows as an example. There are not enough options available to create the window you may require, and yes we can use GDL if your mind works that way or if you don't have much to do and can spend the hours required to make the object you require. We also have Param O but again not sure tif this really helps. We need a fully configurable objects object to cover as much of the stuff the standard library objects don't. Working to deadlines for client means most practices don't have the time to spend on all this other stuff. There are many things in AC which could and should be better and easier to do but currently they are not. Too much time is spent on enhancing BIM in my opinion and very little on making AC the software it could be for everyone. Not just those with GDL knowledge or the larger practices with time to do all this stuff and who work only on BIM required projects. AC have forgotten the small practices who don't need any of this. The New Keynote is an example. lots of users saying its not as good as Cad image or doesn't link to the objects (BIM again) but for small practices this is really a great way of annotating their construction drawings.

Enough of my rant but there is no one size fits all with any software. Each practice works in different ways suiting the work they do and requiring different things from their chosen use software.

[Jim Allen]

You seem to have started Archicad fairly recently because you've been on the forum for a little over a year and you seem to be making a case for Archicad being made to be more like Revit.

Archicad is not Revit. If you want simpler object creation you can use Paramo. You can then develop Paramo objects with GDL script.

Asking for a more-Revit-like object creation tool doesn't necessarily make any sense because it's based on the premise that's the best approach.

Is it?

Your statement "It is beyond absurd we still don't have an equivalent simple and easy method for both 2D and 3D archicad object creation" is also completely wrong.

Making Archicad objects is really easy - you create the geometry, and go to >File>Libraries and Objects>Save selection as object.

Surely that's as easy as it gets?

If you want to create parametric objects you need to put some effort in to learn how to define the relationships. Same goes for parametric complex profiles.

I learned how to build SketchUp dynamic components before I started using Archicad. I learned AutoLisp when I used AutoCad. You need to invest time in learning how to use your tools effectively. The more competent you want to be, the more time you need to invest.

I made the effort to learn how to use Paramo - have you tried it? I have made some useful objects like joist hangers. If you haven't, your criticism is misplaced.

From Paramo I moved to GDL. It was a large investment in time, and I'm not very good at it, but the community here are very helpful and there are lots of very talented and experienced people here that can and will help you. I have made a number of really useful objects, the latest one being a parametric manhole with a custom UI and both 2D and 3D views that can be scheduled, with handles to connect pipes to.

There is also Library Part maker which is arguably easier than Paramo - but that's not parametric.

So you have 4 options for creating Archicad objects, with various degrees of difficulty.

Try Paramo first before piling on the criticism about object creation.

Expecting it to work like Revit though is just silly.

The mesh tool however is crap and well overdue for replacement.

I don't even think there are good enough Grasshopper links to assist with building terrains from AutoCAD meshes, which is what surveyors use. Terrain in Archicad is an absolute pain, and I've worked a lot with it.

We don't even have decent polyline simplification tools to effectively process imported contour lines with thousands of points so that we can build large complex meshes effectively.

And they can't work with coincident z points, so can't be used for walls.

[Kalib St Context]

Again, thanks. I get the points.

I've been using Archicad for the last 10+ years. I train people on how to use it. I've also used Revit for a good chunk of time and I'm very aware of the differences.

My gripes are that we haven't learnt from what Revit does right. Multi-screen, RCP views, global parameters, split levels, grouped items (internal project mod files), sweeps, detail items, filters, model patterns, and live model detail callouts all also come to mind. AC28's big selling point this year was keynotes....Something we on CI select already have....

I've worked with hundreds of archicad users and revit users alike; they all share this library creation frustration with AC.

The family maker may not be the "optimal" solution, but it is easier and more efficient than every other tool at our disposal, as well as has a far shallower learning curve. I can model it via extrusion, set parameters, set formulas, constrain what I need, and be done with it. The same applies to 2D symbols and detail items.

The closest we have got to this is profile modifiers for complex profiles, these have been an amazing help. It also tells me GS is absolutely aware we want more of this. 

I have projects to complete along with many others here. We don't have the time nor the desire to head off and learn code because we need to create some bespoke cabinets where I can adjust the shelves, materials, sizes etc. Ci cabinets solve this, but the principle remains.

The basic way is easy, but designs change, and repeatedly saving objects as static via the save-as is arduous.

We can argue how great GDL and Param-o are all day, but there's a reason we architects are resistant to it, and why most general AC users have no clue how to use either one. I tried Param-o for a while and I am hopeless at it. Same goes for GDL.

Requiring someone to be an expert in coding/grasshopper workflows in a BIM authoring software to expand their day-to-day object libraries is also a ridiculous ask. Most firms lack AC experts, let alone BIM managers that can code and deal with their libraries.

Leave GDL in here for the experts sure, but give us something usable, parametric, and architecturally intuitive for the rest of us to use on the day-day. It doesn't have to be amazing just yet, but some basics would be lovely. Give us the ability to parametrically model new windows, cabinets, fixtures etc.

On top of that, it's about time some new objects are added to the default libraries too.

And agreed. mesh needs a complete overhaul along with the consideration of roads, ramps, vegetation, and paving etc.

[vdentello]

GDL is awesome, yet I agree it's difficult to use and to learn. I have a very hard time teaching.

But more could be done regarding this button, which captures the geometry within the model window.

If there was an option to code with "ready" blocks and not only the auto completing tool, it could enhance the production.

Also, I'd believe this option could simplify the "poly" version captured. It's also painful to substitute every point.