I've been reading through a PDF on my Kindle Scribe which I downloaded from the original site --- it's a great overview, and most importantly, includes citations and references to the various papers where folks shared their research/progress.
> Lemon felt strongly that timesharing systems were the wave of the future and that SDRC should focus on providing software on these systems rather than license its software for a one-time fee. Initially, the company sold ANSYS and NASTRAN on a timesharing basis using computers operated by U. S. Steel.
SDRC did SaaS in the 1970s, before SaaS was cool.
> In September 1994 the company announced that it would be restating its revenues and earning for 1992 through the first half of 1994 to include a $30 million charge relating to sales discrepancies in its Asian operations [...] The company immediately terminated Tony Tolani, a vice president and general manager of SDRC’s Far Eastern Operations.
I was going to say they cooked the books before it was cool, too, but that's a much older trick. They claimed they sold all this software while in reality they just dumped it in some warehouse at the Cincinnati airport. Then proceeded to sell their shares before shit hit fan. SEC link with more details: https://www.sec.gov/files/litigation/litreleases/lr15325.txt
That is to say the downright corrupt and shady sales and management team almost killed the company. It recovered eventually and ended up being bought by EDS
> While the software’s geometric modeling capabilities were improved it was the package’s user interface that really impressed me. Icon menus typically were only two levels deep compared to five and six levels in other systems [...] I wrote in the April 1995 issue of EAReport: “I-DEAS has the best interactive user interface available today for mechanical design and analysis.
I agree. Its interface would look odd and clunky by today's metrics and might even been considered unusual at that time, but it was really nice to use after some practice. They seemingly put some effort into its design.
I remember working with SDRC I-DEAS API in the early 2000s - it was well documented and "clean looking" compared to something like PTC's Pro/TOOLKIT. Interestingly, it was also CORBA-based, and it worked quite well over LAN.
According to the History of CAD Solid Edge originally came from Intergraph. Many packages/kernels seem to have changed hands over the course of decades.
I don't seem to find a dead tree version of the book. I don't know how much work it entails, but maybe the profits of such a sale could go to the good cause they were advertising?
Maybe it's the right opportunity to ask if you know a good online service for printing/assembling of big documents.
>Becoming a book publisher wasn’t on my bingo card when I started Shapr3D. Yet here we go! For now, it’s only available to select customers—but if there’s enough interest, we’ll make it accessible to the public. The book spans 860 pages and weighs approximately 3 kg—just as heavy as CAD is.
Become a select customer of Shapr3D and their CEO István Csanády might send you one.
tangential : I still can't get over the idea that shapr3d license is subscription only. You cannot buy the software. This means that your whole work/workflows ceases to exist if you not continuing to pay. No fallback option.
Unfortunately, the printing industry rolled over for Adobe and set a precedent.
An alternative to Shapr3D might be: https://www.plasticity.xyz/ which has an interesting license and licensing model (I bought a Studio license at launch, but haven't used it since).
Fortunately, FreeCAD is markedly improved, and for folks who want something light-weight and intuitive, I would recommend trying: Dune 3D as discussed here previously:
There's an interesting cultural divide here. Probably a majority of mechanical and electrical designers still think it's risky and weird to use software that doesn't come from a vendor. Even though they experience occasional outages when something goes wrong with the license billing and payment. The electricals are a bit further along, I think because they're closer to the programming world.
I think it’s more that management in this space is very traditional and from a business perspective there’s not much downside to a subscription. The individual engineers don’t really have a choice in cad package anyway and frankly, the alternatives still suck in mechanical space.
Not to mention, file formats between packages are sort of interoperable, but often the design history cannot be transferred. It’s not like software land where if a vendor goes down (and the code is reasonably structured), an organization can mostly replace it with substitutes.
In the open source world there are 2 (maybe 3) geometry kernels that handle NURBS and can produce STEP files. Open Cascade, which is GPL now but had a commercial origin. It's quite good and is used in FreeCAD, Salome, KiCAD (EDA), Horizon EDA, Dune 3d, and probably several more. And then Solvespace which is the only user of it's own bespoke kernel which we are still trying to make more robust. This one is also limited to a certain subset of NURBS constructs but is still quite useful.
Open Cascade may be the most difficult piece of FLOSS IP to recreate outside the Linux Kernel. In fact I'd say it'd be harder to replace given the number of people with interest and the background to work on these two different type of software.
Thank you so much for all the work you and your colleagues do to bring solvespace to the world.
I believe it has the potential to be one of — and probably already is in — a class of super impactful, all time classic pieces of software alongside the likes of Firefox, Gimp, Inkscape, et al.
I've been reading through a PDF on my Kindle Scribe which I downloaded from the original site --- it's a great overview, and most importantly, includes citations and references to the various papers where folks shared their research/progress.
These days of course, one would just get _The Essentials of CAGD_ and similar texts, or read: https://pomax.github.io/bezierinfo/ or watch: https://www.youtube.com/watch?v=jvPPXbo87ds
Highly recommended for anyone who wants an understanding of how modern CAD software came about.
For my part, I have found it invaluable in working on my current project as described at: https://news.ycombinator.com/item?id=43159236
A PDF of the submission? Would be grateful if you had a link; I can't seem to find one on the website.
> Lemon felt strongly that timesharing systems were the wave of the future and that SDRC should focus on providing software on these systems rather than license its software for a one-time fee. Initially, the company sold ANSYS and NASTRAN on a timesharing basis using computers operated by U. S. Steel.
SDRC did SaaS in the 1970s, before SaaS was cool.
> In September 1994 the company announced that it would be restating its revenues and earning for 1992 through the first half of 1994 to include a $30 million charge relating to sales discrepancies in its Asian operations [...] The company immediately terminated Tony Tolani, a vice president and general manager of SDRC’s Far Eastern Operations.
I was going to say they cooked the books before it was cool, too, but that's a much older trick. They claimed they sold all this software while in reality they just dumped it in some warehouse at the Cincinnati airport. Then proceeded to sell their shares before shit hit fan. SEC link with more details: https://www.sec.gov/files/litigation/litreleases/lr15325.txt
That is to say the downright corrupt and shady sales and management team almost killed the company. It recovered eventually and ended up being bought by EDS
> While the software’s geometric modeling capabilities were improved it was the package’s user interface that really impressed me. Icon menus typically were only two levels deep compared to five and six levels in other systems [...] I wrote in the April 1995 issue of EAReport: “I-DEAS has the best interactive user interface available today for mechanical design and analysis.
I agree. Its interface would look odd and clunky by today's metrics and might even been considered unusual at that time, but it was really nice to use after some practice. They seemingly put some effort into its design.
I remember working with SDRC I-DEAS API in the early 2000s - it was well documented and "clean looking" compared to something like PTC's Pro/TOOLKIT. Interestingly, it was also CORBA-based, and it worked quite well over LAN.
Spotted one typo, the reference to Hanchette in the section on Matra Datavision should be Hachette.
That section could also mention that the Matra software is now opensource.
Why are you telling us?
Good counter on John Walker's "The Autodesk File" https://www.fourmilab.ch/autofile/
I recently learned that there is a free community edition of Solid Edge: https://resources.sw.siemens.com/en-US/download-solid-edge-c...
According to the History of CAD Solid Edge originally came from Intergraph. Many packages/kernels seem to have changed hands over the course of decades.
I don't seem to find a dead tree version of the book. I don't know how much work it entails, but maybe the profits of such a sale could go to the good cause they were advertising?
Maybe it's the right opportunity to ask if you know a good online service for printing/assembling of big documents.
>Becoming a book publisher wasn’t on my bingo card when I started Shapr3D. Yet here we go! For now, it’s only available to select customers—but if there’s enough interest, we’ll make it accessible to the public. The book spans 860 pages and weighs approximately 3 kg—just as heavy as CAD is.
Become a select customer of Shapr3D and their CEO István Csanády might send you one.
https://twitter.com/istvan_csanady/status/188829861216722566...
Seeing as David Weisberg's "History of CAD" is trending today I submitted "History of Unigraphics" by 3 of the original Unigraphics 7 dwarfs.
Edit: Unigraphics X SDRC became the system we know today as Siemens NX
It came out at the end of 2024 - these guys must all be around 80 by now.
https://news.ycombinator.com/item?id=43169194
I'll have to work through the 650 pages.
I use CATIA v5 every day, first released in 1998. It's sharp and crusty and huge, but does everything (with the right license).
Funny, I'm French, and I know people who developed a bit of it.
Me too. But I really looking forward to it.
tangential : I still can't get over the idea that shapr3d license is subscription only. You cannot buy the software. This means that your whole work/workflows ceases to exist if you not continuing to pay. No fallback option.
Unfortunately, the printing industry rolled over for Adobe and set a precedent.
An alternative to Shapr3D might be: https://www.plasticity.xyz/ which has an interesting license and licensing model (I bought a Studio license at launch, but haven't used it since).
Fortunately, FreeCAD is markedly improved, and for folks who want something light-weight and intuitive, I would recommend trying: Dune 3D as discussed here previously:
https://news.ycombinator.com/item?id=37979758
https://news.ycombinator.com/item?id=40228068
https://news.ycombinator.com/item?id=40228257
https://news.ycombinator.com/item?id=41975958
There's an interesting cultural divide here. Probably a majority of mechanical and electrical designers still think it's risky and weird to use software that doesn't come from a vendor. Even though they experience occasional outages when something goes wrong with the license billing and payment. The electricals are a bit further along, I think because they're closer to the programming world.
I think it’s more that management in this space is very traditional and from a business perspective there’s not much downside to a subscription. The individual engineers don’t really have a choice in cad package anyway and frankly, the alternatives still suck in mechanical space.
Not to mention, file formats between packages are sort of interoperable, but often the design history cannot be transferred. It’s not like software land where if a vendor goes down (and the code is reasonably structured), an organization can mostly replace it with substitutes.
Is there an overview of which software is using what kernel?
In the open source world there are 2 (maybe 3) geometry kernels that handle NURBS and can produce STEP files. Open Cascade, which is GPL now but had a commercial origin. It's quite good and is used in FreeCAD, Salome, KiCAD (EDA), Horizon EDA, Dune 3d, and probably several more. And then Solvespace which is the only user of it's own bespoke kernel which we are still trying to make more robust. This one is also limited to a certain subset of NURBS constructs but is still quite useful.
Open Cascade may be the most difficult piece of FLOSS IP to recreate outside the Linux Kernel. In fact I'd say it'd be harder to replace given the number of people with interest and the background to work on these two different type of software.
Thank you so much for all the work you and your colleagues do to bring solvespace to the world.
I believe it has the potential to be one of — and probably already is in — a class of super impactful, all time classic pieces of software alongside the likes of Firefox, Gimp, Inkscape, et al.
https://en.wikipedia.org/wiki/Geometric_modeling_kernel
"Kernel developers section"
Thanks!