Thoughts On SolidWorks V6, CGM And Parasolid

For some time now I have been reading comments and opinions from quite a few people regarding the implications of SolidWorks replacing the Parasolid modeling kernel with the CGM kernel from parent Dassault Systemes. Back in October 2010, Matt Lombard wrote a piece about this on his blog titled “The costs and benefits of change“. It’s quite an interesting read. In it he quotes Dan Staples, Development Director at Siemens PLM, mentioning the problems that arose when Solid Edge switched from ACIS to Parasolid. Dan predicts that far worse things will happen to SolidWorks when a switch is made from Parasolid to CGM. Here is a part of the post:

“… changing from Parasolid to V6 is like taking the engine out of your car – the one that the car was designed around – and replacing it with a completely different engine that was never intended to be used in your car. This change is relatively simple if done during the development phase. After a couple of years of selling product it is painful. After more than a decade and within a product that is sprawling in its scope, Dan anticipates with some eagerness that a kernel change will be disastrous for SW users. The change from SW201X to SW V6, considering only the effects of the geometry kernel change (without even considering effects of the cloud) will be disruptive enough for end users that moving to a completely different CAD package may be less disruptive.”

Here is the thing. All these SolidWorks doomsday predictions assume that the Parasolid kernel will be ripped out of SolidWorks and will be replaced by CGM. I don’t believe that will ever happen. Parasolid will continue to be the modeling kernel driving SolidWorks, the Windows desktop version that we use today. I think the current version of SolidWorks will be continued to be developed for some time to come. Maybe “developed” is too strong a word in this case. “Maintained” may be more appropriate. Just like companies continue to maintain legacy products and technologies till they are officially retired. You don’t have to be a rocket scientist to figure out that the bulk of the development effort will be focused on the new SolidWorks which will be using CGM. From the recent noise coming out from SolidWorks I think it is safe to assume that this new SolidWorks V6 (if I may call it that) will be designed in a way that it primarily works off the Cloud, but it will have some offline capability as well.

You need to think of SolidWorks V6 as a completely rewritten new product developed using a totally different set of technologies. This is not going to be like a repeat of Solid Edge’s switch from ACIS to Parasolid. The SolidWorks of today is not going to be ripped apart and chunks of it rebuilt with new technologies. It will continue to chug along like it has been doing for all these years. If users don’t wish to switch to SolidWorks V6 they will be free to stick to the Windows desktop version as long as it is available and be content with it. Most of the goodness will show up on the other side of the fence, where I think the grass will actually be greener. For example, Direct Modeling will most probably never show up on the old side of the fence. And who knows what other goodness the Cloud will bring.

My point is too many horror stories are being cooked up around this kernel switching stuff. Parasolid will continue to drive the Windows desktop version and CGM will drive the Cloud version. It’s as simple as that.

  • Bai Youli

    en,where is the CGM?is there a demo for it?or just somehing living in PPT.

    • CATIA’s modeling kernel is CGM

      • Ab

        Deelip…don’t you think it is premature to talk about solidworks architectures? Is there an official public statement from solidworks declaring a roadmap that can be compared against your suppositions? It is OK to focus intently on CREO because PTC has formally announced, but in this case you sure you really know the plans of solidworks?

        • Hell no. These are merely my thoughts on the subject (see title of post). If this was real SolidWorks information I would definitely not be able to divulge it this way. 😉

          This is one of the main reasons my blog exists. To talk about the happenings in CAD software industry and speculate about its future while maybe analyzing its past.

  • Matt

    Programming isn’t the issue. The issue is data. A V1 user bringing data forward to V6 faces some options:
    – can’t bring it forward
    – comes forward as translated data
    – comes forward as native data
    The first two aren’t acceptable to real users. The last one has to happen without the Parasolid kernel. This is where the difficulty comes. CGM has to rebuild the data the same way that Parasolid did, which is unlikely. This is the problem that Staples was talking about.

    • If you need access to your old data in the native form (feature tree), keep using the Parasolid based SolidWorks. Or maintain a seat just to do that. If you do not then move to SolidWorks V6 and get your old data as dumb solids.

      The problem that Staples had was that he could not offer two solutions and had to merge them both into one. SolidWorks does not have that problem.

      • Matt

        No, it’s exactly the same problem. Old versions of SE still existed. Moving forward means you’re going to lose a lot of data in both cases.

        • Matt, you are not getting the point. Were they going to continue advancing the ACIS version of Solid Edge for years to come in parallel with the Parasolid version? No. They had to get everyone from the ACIS version to the Parasolid version. That does not need to happen here.

          I repeat. If you want your old data in native form, stick to the software where you created it. If not, move to the new platform. You can even do a mix of both if you need to and can afford it. Start new projects on the new platform. Maintain old ones on the old platform. Do whatever works for you.

          • Matt

            You have shown that the situations are the same. The recommendation for Solid Edge users from 199x was the same as your prescription – abandon your live data or stay with the old stuff. What you are not allowing for is that few users will consider leaving their live parametric data behind to be acceptable.

          • And what did the Solid Edge users do back then? 😉

          • Matt

            Bought SolidWorks.

          • Really? And did that solve their problem? Did SolidWorks read their ACIS based Solid Edge data natively? Come on, Matt.

          • Ken

            We updated our models into the Parasolid kernel with little issue and no loss of intelligence. The Solid Edge team did a superb job of translating the features between the two kernels, and aside from a very very small occurence of issues where a feature may have failed, the only issue we encountered was were ACIS allowed zero thickness non-mainifold conditions and Parasolid did not. The conversion was easy too, just open the part in the new version… Done! Parts worked the same way as they had in the previous version.

          • John

            So if it worked for SE, why wouldn’t it work for SolidWorks V6?

          • Ken

            Hard to say, this was done at the 5th released version of Solid Edge, so feature wise, things were much simpler then and you were only working with a couple years of software releases. And, foundationally it was still the same meaning the feature commands still worked identically and all that really changed was something the user didn’t see directly. Not sure it would work as smoothly now as there have been 18 releases since that time and the modeling environment is far more complex today. The Solidworks thing is different in that you are probably changing how the user interacts with the product at the same time, and you will be dealing with 15 years of “history” and added complexity.

      • Deelip, what you are saying is a hard solution too. First, you are essentially saying everyone will have to use two CAD systems, one for new parts one for old parts. Second, this ignores some of the complexities with assemblies and other applications (e.g. FEA, CAM). More and more is done at the assembly level and this means being able to work with all types of data.

        I think Matt is right. The overall issue is handling existing data. I think this post mainly just points out why a “simple” kernel swap may not a viable option.

        • No, I’m not saying that *EVERYONE* will have to use two CAD systems. They can choose between either of them depending upon whether they want to access their old data at the feature and parameter level or not. Back then, Solid Edge didn’t offer that option, which is my whole point.

  • ron

    Autodesk took a different approach when it left ACIS behind: it took the walkway clause in its license with Spatial, and then began rewriting the kernel themselves, renaming it ShapeManger. So, over time, they transitioned their software from Spatial code to Autodesk code, yet kept ACIS compatibility. Gain, no pain!

    • John

      ShapeManager does not maintain compatibilty with newer versions of ACIS. From what I recall, they obtained the source code to ACIS V7 to launch the development of ShapeManager.

      • True. All Autodesk products can save up to ACIS 7 only. They branched off from there.

    • ron

      Yah, that’s what I meant: the ACIS compatibility was not lost from earlier releases of AutoCAD and Inventor.

  • I think all this talk about the “new” SolidWorks is interesting and a lot of people are getting really worked up about it. In my point of view, I have a working version of SolidWorks on my desktop. As long as that continues to do what I want it to do I will stick with it.

    If it turns out that this “new” SolidWorks really is a lot better than the desktop version I will switch to that when the time is right for me.

    However, if it turns out that the existing version becomes outdated and the new version is not living up to it’s hype, I will consider other options. Whether that be Inventor, Pro-E, Solid Edge or something different alltogether I don’t know at this time. But I’m sure I can make that decision when the time comes.

    In my work as a design consultant I already work in a multi-cad environment and although not ideal in terms of data communication I have been able to make this work effectively for the last five years. What I’m getting at, is that it really doesn’t matter much to me which package I use. Being able to create the content I want in a fast efficient way is all that matters to me. Right now SolidWorks is doing that for me, but if that has to change tomorrow, well than so be it.

    • I think you are right Mark and that is probably how it will play out for most existing CAD users.

      A lot of this is likely over analysis as everyone tries to figure what the switch will be to 🙂

    • I think you are right Mark and that is probably how it will play out for most existing CAD users.

      A lot of this is likely over analysis as everyone tries to figure what the switch will be to 🙂

  • Pingback: Tweets that mention Thoughts On SolidWorks V6, CGM And Parasolid | --

  • John

    What if SolidWorks V6 included the Parasolid kernel to handle legacy files. Kind of like the IronCAD “dual kernel” environment.

    Perhaps they have some new whiz bang feature recognition capabilities that can handle 99.99% of the legacy SolidWorks parametric data.

    • Cary

      This topic is interesting. I see it from two points – Data integrity and Command Compatibility. The data integrity is going to be an issue unless they attempt the dual kernel type approach like IronCAD since the model will have to rebuild. However that is not likely since I assume they are making the move to reduce the payments to Siemens. They could do as Deelip suggested but again I assume they will not want to maintain this dual system for long (confusing and a maintenance nightmare). The Command Compatibility will take some time to work out. The kernel will need to be tested to make sure all the calls that SW makes to Parasolid have a mapping in CGM or they will need to rewrite the command (so the comment above that the programming is easy is not necessarily true). The main reason IronCAD uses dual or multiple kernels is to solve modeling conditions that may fail in another kernel (it uses it simultaneously to give a valid result without the user needing to know which kernel is being used) and the ability to share easily with come formats. Really users don’t want to be disrupted in their tasks with this type of switch change (adding multiple kernel support is a better approach). They will have their work cut out for them to make it as seamless as possible or they will lose customer support.

  • Fcsuper

    Of course the owner of parasolid predicts doom and gloom when a competitor stops relying on their product.

  • Anonymous

    Solidworks may be able to do a rewrite with a new innovative geometry engine. It may eventually function as well as the current or past Solidworks version. Their history of geometry bugs and non fixes does not give me confidence. Even now we can close a fully rebuilt assembly and re-open it with the same version of Solidworks and find a rebuild failure. Are they really going to provide the same geometry, I hope so and hope not. I do not like the 1/4 inch errors and geometry failure of the current Solidworks. I really do not like the twisted brocolli shaped sweeps and lofts.

    I must be a control freak. I want control of the geometry. I want to make high level changes and have the power of a relational model to propagate the changes. I want control over my design data.

  • Kevin Quigley

    I get it Deelip. I agree with this and this is something that has been discussed in the past both here and other places. Here’s the thing. CAD is littered with very established systems being phased out as they are bought out or better technology replaces it. I am not talking about small business here either. I am talking about multinational giants switching systems (like for example Mercedes switching to NX from CATIA now). The important factor I think is the geometry and data attached to it rather than the construction history.

    Every release cycle SolidWorks or other CAD vendors has to undergo extensive version beta testing – just to ensure that parts rebuild from version to version properly. Matt and others here have highlighted such issues with SolidWorks, and I have no doubt at all the same issues affect every CAD system that utilises some for of history (I can certainly vouch for having seen issues with VX, Think Design, Cobalt, SolidWork and others). This is part and parcel of moving a point version up. Once we get to SP4 or SP5 most of these issues are resolved.

    If SolidWorks continues to use Parasolid in one version and CATIA in the cloud version I can guarantee that over the course of several releases the CATIA version will be able to read the full SolidWorks data structure and translate this into whatever methodology they are using for user interaction – history, associativity, smart direct face modelling, whatever.

    The pain for the paying user will come after a few releases of the desktop version that has stagnated in functionality while the cloud version takes off (though I have to say if you read most blogs you get the impression all people want is stagnation – personally I pay subs for new features, but there you go).

    My point is that I don’t think this is a big deal. We have, on one hand Siemens pushing SolidEdge/NX ST stuff as the future, we have PTC pushing Creo as I’m not sure what, and we have Dassault/SoldiWorks pushing CATIA v6 functionality (which to be honest I know little about – lend me a license and I’ll be happy to investigate!). The commen factor in all these is that they all say “we can read your native file format, and guess what, construction history doesn’t matter!”

    I think at the moment history does matter, but in a few years time? Who is to say that SolidWorks Cloud will not be able to import a SolidWorks (Parasolid) file and recreate history or smart direct editing features? I think they will.

    When it comes down to it lots of people make lots of noises about switching CAD systems but for most SME’s they will stick with what they have. Why? Because it will cost too much to switch (in capital expenditure and retraining and data migration) and it will be cheaper to stick with what they use. The only valid reason to switch is if the CAD system does not offer the functionality you need to do your job, or if the pricing model radically changes. Anything else is just preference or politics.

    In my career I have switched systems several times – all for the reasons above – functionality and cost. In each case I managed to migrate data without too much difficulty. The only issue I ever faced was when I migrated from Think Design to VX (after Think3 tripled their annual rental charges) and that was because the software stopped working at the end of the rental time.

    There is one other major reason to consider switching systems – when a technology comes along that genuinely changes the way you work. We get a lot of talk of this in CAD but the reality is that most GEE WHIZZ stuff is just that – marketing BS.

    But software companies are clever (regardless of what some say) and from time to time genuine life enhancing productivity tools are released that change the entire market. Look at the way we render now. 5 years ago rendering was a slow time consuming activity that required arcane knowledge akin to a medieval alchemist – they Hypershot came along. Now rendering is something I look forward to. Now almost every major rendering software supplier is developing realtime rendering solutions. That is what I mean about change.

    The same could have been said about the original SolidWorks, or t
    he orginal versions of Ashlar Vellum back in the 80s with the smart cursor. Perhaps if the new SolidWorks brings another step forward all this talk will be forgotten. I hope so, because I think they will, and I hope so because I have my money in the Dassault camp so it is in my interests for this to work.

    • Tom

      ….The only valid reason to switch is if the CAD system does not offer the functionality you need to do your job, or if the pricing model radically changes. Anything else is just preference or politics…

      or performance !!!

    • Tom

      ….The only valid reason to switch is if the CAD system does not offer the functionality you need to do your job, or if the pricing model radically changes. Anything else is just preference or politics…

      or performance !!!

  • Pingback: Bertrand Sicot On SolidWorks V6, CGM And Parasolid |

  • SD

    What I wonder about: I´m only hearing doubts about the future development of SolidWorks.

    Remember SolidWorks brought us fantastic functions an possibilities over the years, leading cad development at all. So, why worry? I would worry much more about my subs if they did not restructure to provide even better, faster, intelligenter and more useful software. That´s the main reason why we pay them from year to year: making them think about getting better. I do trust them to bring more power to me. Because they did in the past.

    Yes, I am aware of rebuild errors, system crashes, translation problems etc. – but can you name any software that does better than SolidWorks. Problems are the same everywhere and it seem´s to me SolidWorks is the one-eyed between the blinds.

  • Guest

    The fact that over 92% of our customers renewed their SW subscription in 2010 is strong evidence that people have confidence in SW.

    • The 92% must be a number for the US. I work with a variety of vendors in Taiwan and China. Although they all own legal copies of SolidWorks only 1 of them is using SolidWorks 2010. All the others are using versions anywhere from 2004 to 2010.

      Given the fact that I work with a variety of vendors that are not on subscription the 92% number seems very high to me.

  • Neckie

    The fact that ‘over 92%’ subscription renewal rate is rather standard at most CAD vendors is a testament to the acceptance and effectiveness of subscription-driven business models rather than confidence in SW in particular.

  • Billy Oliver

    The issue is data conversion between kernels. And what percentage of the data converted will be error free, 100% data conversion?? SW has 15 years of user data that utilized the parasolid kernel. No one will risk their data converting legacy files to a new kernel if they cannot guarantee 100% converstion rate.

    Looks to me that SW is going to the Catia kernel. And that they will say that the conversion of legacy data will at first be low but in a few years they will have a 100% conversion rate.

    But I think that it will be much like the SDRC and UGS merger went. They said 3 years until 100% data conversion for the SDRC users moving to NX, previous UG customers (parasolid) did not have the conversion problems that the SDRC customers did.. Well its a decade later and there are still SDRC users out there who have not converted their legacy data. What has happened to the SDRC user’s, their software has “stagnated”. Bugs have been fixed but all of the “evolutionary software development” has gone to NX. Now with ST they are saying that the SDRC users can finally convert data 100%.
    So I think the SW to Catia kernel will go much the same way. The new SW V6 product line will get all of the “evolutionary development” with the direct editing stuff etc. The SW ParaSolid (SWPS) product will “stagnate” but have to be maintained to support the 15 years of the SWPS parasolid legacy that will not be open able in SW V6. I have heard that the SW – UG parasolid license agreement is set for the next several years. So there will be a SWPS product for a while.
    It all about the legacy data. Remember the guys who started PTC were old ComputerVision and Applicon guys. They showed the former companies the new fangled “parametric solid modeling”. Both companies flatly rejected the “feature based” solid model because they could NOT “obsolete” all of their customers 3-D wireframe data over night. Well you know the “rest of the story” on that.
    Customers have changed from CAD to CAD throughout the years, for various reasons. SW has had a 15 year run of stability. If it is true that they are about to undertake this change, I believe that a “large” chunk of their customers will switch to another “parasolid” based software pack. Which one? Well looks like NX or Solidedge to me.

    The only thing that surprises me in this Catia V6 kernel switch is. Why did they take so long to change. I think they might make it a Catia Light. Then abandon the SW product eventually. But who knows.

  • I think Dassault will assume the same position that they have whit Catia, you will lose all your model information(history tree) but you will be available to import it to the new version. This sucks for a lot of existing costumers and will open the door to possible incumbents. But this is the Dassault way of moving forward.

  • Why not to develop neutral feature format for CAD of ISO standard? Isn’t this problem is faced by every ISV at some stage ?

    Let the bigger good come out of this. Agreed this is mammoth task but not the impossible one. In the absence of it has created so many companies to work on CAD interoperability.

    One can argue STEP did not help, how new format would be ? But haven’t STEP helped to improve translation as compared to IGES or DXF(DXF was one of the defacto for CAD translation) by many fold.

    Non willingness’s of CAD industry to collaborate on this has hammered end user by infinite year of human efforts and uncountable monetary loss.

    For CAE simulation many user still prefers to translate geometry in crude STL format. As every CAD system writes STL file accurately which every other CAD/CAM/CAE/CFD system reads properly. What a waste of rich data… Things have changed recently still it has not gone ahead of STEP and IGES.

    Just because no one is willing to make efforts on standardization ?
    It’s not that they can not but is more of they don’t want to.

  • Guest

    Much of the discussions here  is about the migration and how seamless the transition should be from Solidwork 201x to Solidworks V6.

    I don’t hear too much about the limitation of current kernel.

    The reason for the new kernel may be solving the 1 km limitation on part.

    This limitation really stop me right on the track.

    I would like to incorporate large object such as mountain contours, do mega plant layout.

    With current Solidworks kernel,  it is too limited to do that kind job.

    • Ken

      But it was also never meant to do that job.  That’s what Plant Design software is for, and you wouldn’t expect it to be any good at mechanical design…