Tried to dabble in AI for parking layout. Started out ok - but then I threw it an irregular polygon. It thought about it. Said this is hard. Then said pay me money.
The way these vaporware firms are going I think giving a few grads the parking standard will be a more cost beneficial solution than investing in these types of Revit hack apps.
Most of them in my xp can operate in just their own sample models.
You got me down a rabbit hole of thinking/complaints (/rant?)
Not only you to be fair, but same day seeing this https://discourse.pyrevitlabs.io/t/revit-python-shell-update-to-revit-2025-revit-net-api
I feel the OSS world rely only on single person will to do stuff and share.
Many years back, I backed something around parking lot and industrial building iterations with Dynamo and Gen Design module.
At the time, TestFit was close enough, but my client wanted me to try and make something similar in dynamo. Pretty tedious experience. The end product was OK / close to TestFit even. But the gen design module was so slow and unreliable that we tossed the thing in the end.
I feel this is the kind of project that is better handled working with software companies directly than trying to make yourself in your garage. The amount of research, work, debugging, and edge cases management implied is huge and people tend to neglect it.
Software dev takes both time and money. I feel that the approach of making smaller scope tools is always more reliable and makes more sense in the long run that trying to fix these complex tasks with one man show teams (unless you are so brilliant, and LLMs are so perfect that you built these types of engines overnight !)
Something like this tool, upcoming very soon, Full models audit by jmcouffin Ā· Pull Request #2503 Ā· pyrevitlabs/pyRevit Ā· GitHub
took me weeks to design and code, becauseā¦ shitty Revit API, edge cases, bugs, bad coding , bad refactoring, too much refactoring to make it clean, ā¦
Making the same code closed source/C# and paid, I doubt it would pay itself in a year if I put it on the autodesk app store. Whatever the value of such a tool.
Again, AECO as an industry is OK (up to a point) paying 3000USD a year for their design tools, very happy to use pyRevit and end of the day, not making any donation for us to be able to sustain pyRevit properly and not like poor lads. _ we got like 5 firms who donated less than 100 and BIG who donated 1200.
I wish we could have 10x the donation and hire 1 C# dev to make pyRevit less buggy. :dreaming:
Less Issues, Less complaints, More fun creating cool tools (because this is the fun part)
drop the
picks up the
When it comes to supporting pyRevit specifically fiscally, the one thing thatās stopped me from going to my company and bringing up supporting this project is what deliverable will they get out of it? The opencollective page is nice, but right now it seems weāre just throwing $ in a pot without knowing how its going to be spent.
The big desire is of course getting a full time maintainer, or hire a dev just to get v5 out the door. If a dollar amount could be put to that, then itād make āthe sellā much easier.
Folks at work are asking my why there isnāt pyRevit in 2025 yet - and every time I explain how its OSS & relies on volunteers. If there is a dollar amount to get v5 released, then Iād think more donations would come in.
puts the back on the mic stand
I like that you picked the mic!
Agreed, and right, nothing has been spent since we opened it.
The amount available and the recurring donations arenāt enough to āhireā seriously someone (roughly 5000 in the post and 65$ recurringā¦)
We should advertise the plan, but the thing is, āweā is mostly āIā + @sanzoghenzo and @dosymep in terms of involvement, and while there is a plan to refactor the code base core of pyRevit into C#, it is not advertised.
pyRevit WIP installers are 2025 friendly.
I will push the button of the release probably next week, I was waiting to finish my full audit tool and fix the certificate. The first is close to down, the second was finished yesterday.
All that being said, I will try to communicate more on the grand scheme of pyRevit, then you will have the bullets to make the sell.
That being said, this still bother me, as I candidly think that pyRevit brings a lot of value to AECO:
Just so I donāt incorrectly get too excited, are you saying v5 should be coming out next week (obligatory āif all goes to plan and whatnotā)
Iām not familar w/ the world of freelance software development. In terms of a dollar amount would this be an entire salaryās worth or a half, third, quarter of it? Not sure how many hours/week would be needed of a maintainer.
Thanks, I appreciate it.
100%, mainly just when it comes to time saved on scripts. That one of the reasons I want to implement telemetry so I can produce a āwe saved xx billable hours this yearā
Weāre getting off topic on AI parking, but thanks for the response
yes
Iād consider having someone that can put letās say a full week worth of C# magic would be a very good base (every month)
There is the knowledge gap to fill at first obviously, not every programmer knows the tech stack and The Revit API (plsu all the pythonnet ā¦)
no worry, the mods wonāt ban you
Back to parking (which we do way too much of with retail layout). It isnāt TestFit.io level by any means, but it is ok.
Weāve got a pretty good manual set of tools for parking lot layout.
- A series of floor types with the different parking patterns as model patterns. This makes it simple to establish a quick visual of the layout, spacing by just pushing the pattern around.
- Outside rows and drive isles are just a floor with an offset from the main field.
- All of the parking families are line based and quite robust. They can be tagged for count of each row and report different parking field totals. Being able to use pick lines is sweet.
- We have a random car placement and coloring tool to fill in detail for rendering in VRay , by swapping out cars with VRay proxies in 3ds Max.
- The next step is to just automate the construction of the islands, road striping, plantings, etc.
Right now, Iām using the floor edges and the floorās pattern to generate the placement of the line-based parking families. The plan is to link these elements with data so I can update the floorās and just regenerate the layout. Also need to work up the parking families to be able to comply with zoning - such as no more than 12 cars in a row and calculate percentage of green area.
From my limited experience, open source projects work better if there is a company behind it, or when a long-sighted company uses it and allocates employees time to make it greater.
This works well in sectors that have let go of the concept that everything made in the work hours should be covered by intellectual property, something that the AECO sector is still reluctant to do.
A random c# developer without knowledge of Revit, its API,the nature of the project and AECO in general is not worth much, maybe less than a smart engineer/architect with little programming knoledge and time to spend learning and developing in c#
100%
This is rant day for me, but things need to be written for concepts to be conveyed
Sorry for the parkade ai conversation
ā¦at this point in my life, Iād love to step out and start an open-source BIM software platform written on top of Linux.
Just to get the whole parameter Tower of Babble straightened out. If I could just get my ex-ADSK developer to go along with the idea.
That what you get after a tough week in AECOā¦
Rants, and wishful dreams ā¦ And no parking lot at the horizon
I will also missuse this conversation. Did you contact ricaun ricaun | LinkedIn ? The company produces revit plugins in C# as well as contributions to OSS. Seems that the profile corresponds to what pyRevit needs in terms of C# work.
Yes I have,
No reply, then he commented jsut last friday on something I wrote, I wrote back, nothingā¦
Letās keep it going. Itās free parkingā¦
Might need to consider a company that is committed to this, such as a DiRoots type. I appreciate itās not in the open spirit of things, but Iām unsure you will get ongoing C# support from people that arenāt involved in pyrevit as is to any better degree than a firm that takes on these tasks.
Cost is probably a barrier thoughā¦
Roman would have been my recommendation as well, been using their templates. Not sure of their Py skills though, maybe Ehsan has some connections?