Next Generation SolidWorks to be Direct Edit
An article on Graphic Speak blog by Randall Newton contains an interview with Gian Paolo Bassi, the new VP of R&D at SolidWorks. One item of interest in the article comes when Mr. Bassi is asked about his vision for SolidWorks:
GfxS: What is your vision for SolidWorks the product?
GPB: We will focus on performance, and not just the pure architecture of the product. The future of SolidWorks is about providing more power, more ability, more tools for collaboration, and the opportunity for exploiting smart phones and tablets. We are excited about the opportunities to improve performance.
One key to improving performance is to rid ourselves of the history-based design paradigm. There has been a rediscovery in the industry of a need for direct manipulation of shape, to give designers more freedom. We will work to give SolidWorks better design freedom performance, better computational performance, and greater flexibility…
The answer or the question is possibly not as defined as it might be. Before you can consider the content of the answer, I think you have to clarify that even though the question clearly is talking about “SolidWorks”, the answer is about a product that does not as yet have a name, and has been called variously “SolidWorks V6”, or “Catia Lite” for lack of a real name. So this is not really about what we think of as SolidWorks, it’s a brand new product from Dassault.
What I’ve read in other places is that the Catia V6 platform is said to be based on what they are calling “declarative modeling” . At this point, any distinguishing marks between “declarative” and “direct” are opaque, but I don’t see any mention of combining direct modeling with history based modeling, in the way that Synch Tech does.
What, if anything, does this mean?
First, it means that transitioning from SolidWorks 2013 to Catia Lite 2013 won’t be the utter feature failure nightmare that you imagine when changing kernels. That’s because there won’t be features in the way you’re used to thinking about them. Direct edit (and I suppose declarative modeling) works with what history users would call “dumb solids”. They may arrange for some sort of “features” like the features in Synch Tech, where … wait.
It looks like the next generation SolidWorks is going to work much like how Synchronous Technology has worked for the past 4-5 years, but possibly without the flexibility of using history modeling where it makes sense.
So. If you plan on following the SolidWorks directed path and moving beyond the current SolidWorks product, which becomes obsolete as soon as Catia Lite starts shipping in 2013, it will be pretty much just like buying a new CAD package and starting over. New interface. New paradigm. New pricing. New product names. Everything is new. So what should you do, move to a brand new product which might take 3 years to work the bugs out, or move to an established product? Siemens (Synch Tech), PTC (Creo), and even Autodesk (Inventor Fusion) already have software in the marketplace that have different methods of combining direct and history modeling. SolidWorks (or shall we just call it Dassault, because the SolidWorks product remember will be phased out at some point after Catia Lite hits, and the SolidWorks name is not even on the new headquarters building) Dassault, for all their chest thumping about doing this so that they are first with new technology, will wind up being dead last of the four major players. Their silence on the issue has been meant to buy them time so that customers aren’t thinking too much about direct edit, or that history modeling has some limitations.
So does Bassi’s description sound good? Yeah, it sounds like there are some useful ideas in there. A lot depends on specific implementation, and anything can happen, but it’s potentially a big improvement. My question is, how are they going to induce you to move to Catia Lite without causing you to consider Synch Tech, Creo, or Fusion?
Notice the only mention of “cloud” in Randall’s article is in the author’s caption on an n!Fuze screen capture. Does that mean they have backed off full blown CAD in the cloud, or they are just being coy? Why would he not ask Bassi about cloud, since cloud is one of Bassi’s “things”? Bassi was previously at a company called Riwebb, which had this on their site:
The site has since been taken down, but Bassi did CAD and “web enabled solutions”.
No inherent roadblock, other than architecture.
Think about it this way: Say you built a house 20 years ago. Every year, you’ve remodeled and added on to the house. It now has 5 basements, 27 bathrooms, 102 bedrooms, is 47 stories tall, and is made out of combination of bricks, straw, steel, and bailing wire. And now, you want to add climate control, so that each room can be kept at its own temperature. Without having to move out during the process. Adding functional modeling to a CAD system might be slightly harder than that.
@Evan Yares
But would it be safe to say that there is no inherent roadblock to implementing functional modeling in either history or direct modeling systems?
“In computer science, declarative programming is a programming paradigm that expresses the logic of a computation without describing its control flow.” That’s what Wikipedia says, at least.
Declarative and functional programming are closely related. For practical purposes, I’d say the terms declarative and functional, in reference to modeling, are pretty synonymous. They both relate to the functional characteristics of features, rather than their purely geometric and topological characteristics. So, functional/declarative modeling is an attempt to make the system work in a way that is closer to what you’d hope for. (So you spend less time saying “I can understand why the system did that… but it’s still stupid.”)
Direct modeling systems have some inherent (but limited) functional modeling capabilities, in that they need to infer features (and functional intent) in the process of editing.
@matt
It’s pretty much the same thing, except that DS (and Bassi’s team) has had 10 more years on the development of it. I’m out to see the DS guys to have a good dig into V6 (something that has taken 6 or so years to finally arrange), so I’m curious to see exactly what V6 can do live and not just in a vague demo or video.
Regarding Functional vs Declarative, I think its just a question of naming conventions for the same thing – I know I had trouble describing what the ImpactXoft stuff did at the time because it’s so subtle. Maybe it all boils down to an feature-based modeller than adapts to your requirements. Inventor has elements of it, SolidWorks does, V6 certainly does, but so does IronCAD and Siemens’ tools.
Al
@Al Dean
Al, I got a look at ImpactXoft in 2002 or 2003 when they were trying to set up resellers. The closest history analogy seemed to be that some features had “gravity”, and would stick at the bottom of the tree, and new features would get stacked on top of that feature. But that is just what sheet metal features do in SolidWorks. And as I recall it was called “functional modeling”. Not sure about the relationship between declarative and functional.
The Catia video link you gave shows functional modeling, and the very first feature they create works like the SolidWorks Indent feature with a tool body and creating an indented wall thickness in the plastic housing. The SW Indent feature is entirely history based from the end user point of view. It’s also one of the coolest (and most underused) features in SolidWorks.
In fact, I linked to that same Catia V6 video almost a year ago in a blog post about declarative modeling: http://www.dezignstuff.com/blog/?p=343
There was at one point a working alpha of a plastics environment that worked like sheet metal and weldments, with those “gravity” features, or features that establish “specifications” for the rest of the model, like the thickness or bend radius in sheet metal, except for plastics type of features. Never turned up in released software. Would have been very cool for plastics design, which I do a fair amount of.
Can you describe any difference between “declarative” and “functional”?
Here’s a link that adds to the confusion: http://perspectives.3ds.com/interviews/direct-modeling-with-dassault-systemes/
You can have functional modeling in a history based system, I don’t see any conflict there. SolidWorks has had several examples of this for some time, with the Mounting Boss, Vent, Lip/Groove, and Snap Hook/Groove features in addition to Indent. These features came to SW after the IX acquisition. They definitely stand apart from the typical extrude, revolve, sweep type of features, because they don’t involve the typical [sketch]==>[geometrical process] workflow.
The Inventor clip is all stuff you can do in SW with the Indent tool. At least the first half of the movie, the video cuts out for me at 3:45.
Thanks for the background.
Declarative modelling is not direct editing. Let’s not get the two mixed up because at a very fundamental level, they’re different. Declarative modelling is something that Bassi has a heavy background in, having worked with Attilio Rimoldi on the technology at ImpactXoft and subsequently DS (It’s inside Catia). If you’re not familiar with ImpactXoft, for those working with plastic part design, it’s worth a little look. The background is a co-development project between Rimoldi’s outfit and Toyota. Rimoldi took the technology forward with ImpactXoft and Toyota Caelum XXen (www.caelum.co.jp).
The first time I saw it (implemented in Caelum XXen), I couldn’t figure it out. It initially seems to be a standard feature+history modelling technique, but when you see it in action and play with it seems to defy the laws of history-based modelling.
The reason it’s called Declarative Modelling is centred on the idea that you ‘declare’ features and function that you want from your model and subsequent features have to respect it. Examples are quite hard to pin down with text, but think of this.
a) Take a solid model.
b) Shell it
c) Now, create a sketch and ‘cut’ into that shelled part.
What happens now?
In SolidWorks (unless you use the deform feature): The system will cut out the geometry and you’ll have a gaping hole in the part
With declarative modelling: you can get a recess with the shell adapting to the geometry or a cut if you flipped that particular switch in the software.
That’s the basis of the whole thing and the freedom from ‘history’ based modelling. You ‘declare’ the functionality or feature you want in your part (in this simple example, that its thin walled) and the system works its arse off to maintain that. Obviously it can’t create the impossible and it can be very very subtle, but for anyone with experience of a history modeller, watch it in action, and it’ll make your head spin if you concentrate.
There’s a nifty DS video on how its used in Catia V6 – (http://www.youtube.com/watch?v=yqaalrAPij0)
Inventor users might recognise the same tool within Inventor. Yup. Guess who developed that too. Same team. (http://www.youtube.com/watch?v=4amBdN8tXSE)
Hope that helps
Al
@Normand C.
Your information is slightly out of date. True, we have had the “fold up” of 2D to 3D for many years, but in ST3 (Oct 2010) we added the ability to take a SolidWorks part and its companion drawing (in dwg) and inject the dimensions from the drawing into the 3D solid model and then drive it. You can see this in the demo called “Transfer design intent from 2D to 3D” or something like that. The third entry in the Synchronous demos found at:
http://www.plm.automation.siemens.com/en_us/products/velocity/solidedge/demos.shtml
I wonder if Dassault understands the gravity of the situation. They’re making a new product that will compete with their own existing one. If for any reason this new product is not better than the existing one it could easily be the end of Dassault. The acquisitions that Dassault have made lately are very questionable and don’t address what the CAD industry is in need of. There is a huge amount of talent at Dassault, unfortunately none of it is in management. I just hope they haven’t hit a brick wall with this new product. How long have they been working on it (3-5 years)?
Whatever the name, the solution or the road, SW is now behind in this new trend of direct design.
With that face will say now that the parametric is still the best or only solution?.
They need to swallow its pride and change.
@ Jeff H.
I agree fully with your last sentence.
As smart as we all think we are (myself included), I doubt that any of us can accurately judge the outcome of a product that is not fully defined yet. I see benefits from Direct Modeling but I also see strength in History-based models that require some forethought. I understand the pitfalls of “design by committee” but I would be scared of any product that was developed solely by myself or any 1 other person. I appreciate everyone’s right to their opinions but I doubt that any statement similar to “SWorks is dead” can do anything but lead to skepticism.
Actually, I think Mr. Bassi mentioned “cloud” without mentioning “cloud”. In the quote you provided, he states, “…We will work to give SolidWorks better design freedom performance, better computational performance, and greater flexibility…”
—————————————-
GPB: “…We will work to give SolidWorks better design freedom performance…”
My Interpretation: As soon as we release our cloud software (whatever we decide to call it and whenever we decide to release it), you have the freedom to discontinue using SolidWorks because it will no longer exist. As soon as our cloud software liberates your computers of your design data, they will perform much better. After we force you to move to the cloud, you have the freedom to use other CAD software. However, good luck getting your design data from us. Remember – because we are part of the Dassault “family”, we can afford more lawyers than you. :-r
—————————————-
GPB: “…better computational performance…”
My Interpretation: Your processor(s) suck! Our cloud has more cores than you. 😛
—————————————-
GPB: “…and greater flexibility…”
My Interpretation: You will finally be able to use SolidWorks on a Mac, or any other operating system, computer, or mobile device…..NOT! No, but seriously, you will still be able to experience the frustration of all the crashes and bugs of our new cloud software on any OS or computer/device. That will be sweeeeeeeeeet.
@matt
It’s actually not a way to reattach a drawing to a 3D model, it’s a wizard to build a 3D model out of 2D views. It helps you place the drawing views on specific planes in a 3D model. It’s been in Solid Edge for many years before the advent of ST.
So to answer Jason: ain’t working. SE can open SW parts and assemblies, but not .slddwg files. To be able to open SW drawings in SE, you’d need to export them to DWG/DXF, thus loosing any associativity. As for assemblies, all links are lost. So that’s dumb solids, and dumb assemblies.
In case of a CAD system switch, I believe the only reasonable alternative is to keep an installed copy of SW for the legacy files. But that’s the same for any CAD package.
“how many hours of lost productivity can companies afford to lose to being multi year beta testing program guinea pigs?”
Absolutely,.. Cui bono?… this is a very serious and real issue ALL SolidWorks Users should be extremely concerned with!
BTW,.. the comment by this R&D guy about being innovative… were we born yesterday… or is he smoking crack!?
Hmmmm. not sure what to think of this. I spent a year and a half using CoCreate and dumb models were a good way to describe things. The models were so dumb that it would forget you created a fillet when you tried to remove it. Direct editing is great for simple geometry but when you start talking complex castings forget it. You where better off starting a new model from scratch than trying to update something. The Creo approach made the most scense to me because you could (according the the propoganda) switch back and forth no problem. I guess we’ll just have to wait and see what happens.
Carlos,
Another couple of things to think about. Siemens writes and owns Parasolid and there are things in there that make direct editing, the Synch tech variety, work really well and they do not sell this to others. So Dassault has this impediment to deal with if they try to do this on the Parasolid kernal. I wonder just how this Catia Lite will play together with Acis, Parasolid, history and non history stuff and reliably import other file formats and accomplish a high degree of accuracy and reliability without years of refining. Yes, the last of the big four to see the value of direct editing and the first to accompany it with a kernal change at the same time.
The real question Chris is how many hours of lost productivity can companies afford to lose to being multi year beta testing program guinea pigs? I went through this with SE where the first two versions were a lot better than not having direct editing but I would not have recommended ST without reservations until ST3 and this was without a kernal change thrown in the mix. I can’t imagine what it would have been like with the two combined. One thing is for sure, we will all know but the wise thing to do would be a spectator and not a participant.
@matt
Matt,
Chris may not be so far off if you consider what AutoDesk did with Inventor. They were late to the game with an MCAD product (don’t count mechanical desktop) and then they bundled Inventor with upgrades/subscription with AutoCAD. Then they claimed the number 1 selling MCAD software. It seemed to work ok for them…
SolidWorks using Kinect !
When will see it ? Aren’t the product managers in SW dreaming of using the video controllers to drive design ?
One thing I’d like to see is an overhaul of the SolidWorks Business Model, in these areas;
Subscription Maintenance
Service Packs
Software Testing
Bug Reporting
Help Documentation
Resellers/VARs
Commission CAD Salesmen
I think in a nutshell the current SolidWorks Business Model is outdated.
Matt, I’d like to hear what you have to say about this, thanks.
Devon Sowell
@Jason
Solid Edge has a way to reattach imported drawings to solid models. I haven’t played with this much, but it looks cool. http://www.youtube.com/watch?v=sd9g0asl_j0
What about legacy data. I know direct editing modelers can edit models easy enough but what about the links to assemblies and drawings. If I open a model and drawing in another system like Solidedge, can I update the model and the drawing update? If I have to re-detail a lot of drawings, that’s a lot of work.
>The future of Solidworks is about providing more….
Solidworks doesn’t have a future.
SWv6 is about providing ‘enough’ without impacting Catia sales.
If they spent the last 6 years bringing SWv6 up to where SW is today then it is doesnt have much of a future ahead of it. I still think this is about CAD serfdom rather than anything.
Adding the collaborative and social media bits is a one off, the kernel change is a one off, get these things out of the way and your life is going nowhere fast.
Remember a lot of the new capability is already available to you via Skype or Dropbox, or was via TsElements.
So far DS haven’t answered one question about the detail of how this is going to work in practice, hell it doesn’t even have a name yet.
If there is one characteristic about SW that absolutely pisses me off it is the way they can’t front their customers. That goes from missing documentation right up to CAD gone missing.
Frankly I would rather talk about any other program than persist with the DS vision of the future.
Having said that I am going to cut back on this blog I’ve decided I’m going to confine my comments here to articles about SE or whatever else of interest happens along. 😉
I was skeptical about Mr. Bassi in the beginning, but after reading that article I have changed my opinion. I am very curious about the price of the software. If they do want to maintain current customers, they will have to have an upgrade pricing model that will be less than buying some other vendor. GPB did admit that a lot of mistakes were made with nFuze!, one being the pricing model.
@Chris
Really? ;o)
I believe every SW user should ask themselves:
“Should I trust my company data to a brand new, unproven technology, or should I just go for the trusted and tested Synch Tech that is on their 8th. year of development?”
I know what I would answer…
Note: Synch Tech development started in 2004 but was made public only on 2008.
“”My question is, how are they going to induce you to move to Catia Lite without causing you to consider Synch Tech, Creo, or Fusion?””
Matt, I think just bye money.. When you move to Siemens, Creo or Autodesk than you have to buy software. But when you move from Solidworks to “catia lite” then it will be for free, just pay the maintenance further.