2024 Technology Preview Program:
Master powerful new features and shape the latest BIM-enabled innovations
2020-08-19
10:14 AM
- last edited on
2023-04-12
12:41 PM
by
Laszlo Simon
2020-08-19 12:19 PM
2020-08-19 01:29 PM
agroni wrote:It depends. Programming languages are designed for different purposes, and even their implementation may have specific purposes in mind. Python is considered 'easy' because you can get almost instant results for minor problems, but it doesn't make hard problems easy – it only moves the core of the problem somewhere else, which can ultimately make it much harder to solve. I wrote a GDL object some years back – Modular Joinery – that would have been vastly easier in C++ than GDL. The simplicity of GDL means that it is only appropriate for simple problems, and in fact that is precisely what GS has designed it for. Roughly 80% of our bugs come from GDL for the same reason. It's acceptance of almost anything is a double-edged sword. I've seen very large firms dig themselves into terrible IT problems by the application of inappropriate solutions, e.g. trying to use GDL as a tool or a spreadsheet as a database. The same is true for most 'simple' scripting languages. So consider – what kind of problem do you intend to solve? GDL for 2D/3D illustration, python for external batch processing, C++ for tight integration or larger-scale applications/add-ons. If you want the best toolkit, learn and use
I will post this general question here as I was wondering for the future, what would you prefer to learn as a programing languange and implement it in the Archicad environment?
2020-08-19 09:11 PM
2020-08-20 12:17 AM - edited 2023-01-24 09:42 PM
agroni wrote:
Thanx guys for your detailed answers
To be honest I am a tabula rasa in programing, but seeking new horizons to improve quality and efficiency in our office.
I can observe some repeating routines that my friends are practicing in our office that I believe could be more automated and save tons of working ours and reduce mistakes. We already have many things in Archicad implemented but we have reached the limit of the programe itself, hence it is time to expand the implementation field.
I am talking about defining alphanumerical values for spaces or even components. One good example (which I belive was made in Python) is from the office Aidea about automatically generating layouts and placing drawings on them.
Hi Mr. Agroni, could You expand on the limits of the software (archicad) that You have reached? What it is that you want to do that it can't do? (Other than truly automatic dimensions, of course, which should be a thing by now)
Also, for anyone with this kind of experience here, at what project size would it be recomended to try to implement this kind of automation?
This brings again a question we have been discussing before: Why should i, as an architect, learn to code in C or python? (i know Mr. Moonlight, as a very advancer user, is a big advocate of this). I mean, im totally into having better processes and stuff, but why can't i just pay for this add ons if i ever need them instead of putting years in learning to code? At what point is it better to do it in house?
We have survived so far without it, why now? what is the conceptual justification for this (other than the typical "everyone, eventhough i dont know who, is doing it and the ones that dont hop into the train or will be left behind")?. Some of the offices that advocate for this "automation" churn out generic boxes that they want to sell as architecture. Im not sure if we should go down that path that easily.
2020-08-20 05:57 PM
jl_lt wrote:This is a general question today of how the profession of an architect is shaping. Thirty years ago we drew on paper, today we have digitized those lines... So is programing zhe next step in this profession? It's to early to tell... Maybe... Who knows...
Why should i, as an architect, learn to code in C or python? (i know Mr. Moonlight, as a very advancer user, is a big advocate of this). I mean, im totally into having better processes and stuff, but why can't i just pay for this add ons if i ever need them instead of putting years in learning to code?
We have survived so far without it, why now? what is the conceptual justification for this (other than everyone is doing it and the ones that dont hop into the train will be left behind)?. Some of the offices that advocate for this "automation" churn out generic boxes that they want to pass as architecture. Im not sure if we should go down that path.
2020-08-21 10:29 PM
2020-08-23 05:13 AM
2020-08-23 04:54 PM