Some Thoughts On CAD On The Cloud

My keynote speech on the future of CAD at the Bricsys 2010 Conference in Bruges, Belgium, led to a few interesting discussions at the conference and comments on this blog. One person asked me jokingly, “Do you say all these things just to provoke people?” Come to think of it, I don’t think he was joking after all. 😉

At the outset, no matter what you think about me and my views, I must tell you that I am as uncomfortable as most people about CAD moving to the Cloud. However, I believe that the future almost never looks anything like the present. If it were then we would still be using drafting boards and writing letters to each other instead of emails. Also, my observations/views/predictions are not based on any confidential information that I have received from any CAD vendor regarding what they are doing with the Cloud. They are purely based upon how CAD vendors have acted in the past, are acting at present and will/may continue to act in the future.

One of the things I mentioned in my speech was the versionless face of CAD in the Cloud. I pointed that that there was only one version of Gmail and Google Docs, the version currently running on the Google servers. I got a few comments about that.

Bob Mayer of IMSI/Design commented: “Deelip – not sure that I completely agree with the notion that computing on the cloud will be ‘versionless’. While subscribers will always have the most current version of the application, the models and drawings created over time will change in structure as the application evolves. So, in this sense developers of cloud applications will need to insure backwards compatibility with the content previously created. And, of course, there will be continued interoperability issues with consumers of this cloud-created content, when it is used by other desktop CAD apps., as well as other cloud-based cloud apps.”

Here is the thing. In cloud computing, the data resides in the cloud where the computing occurs. So when the software is updated the vendor will automatically port all the old data created by users to the new format. This is what happens in Google Docs, Facebook, Twitter and everything that happens in the cloud. Do you even know or care which version of Facebook you are running? If a user imports some old offline data that he created long ago into the cloud, that data will be converted to the latest version.

Dietmar Rudolph didn’t quite like that scenario. He commented, “… and of course they ensure that such an automatic conversion will keep your data 100% intact, that no outdated objects are lost and that approved designs still are approved.”

Frankly, I don’t see how this is very different from what CAD vendors do today with their offline software. When you open an AutoCAD 2007 DWG file into AutoCAD and save to 2010, AutoCAD does the needful and migrates the old data to the new way that it has been defined. The same thing will happen in the cloud as well. Its the same thing and will work the same way when it happens in the Cloud.

As regards interoperability issues with users of offline CAD, this will just be another way to force people to move to the Cloud. Pretty much how all the suppliers of an OEM keep up when the OEM moves to the latest version. That solves another big problem for the CAD vendors, which I somehow forgot to mention in my presentation. There are suppliers who pay through their nose to maintain a seat of CATIA not because they want to, but because they have to, so that they can work with a company that they supply to or work with. If and when that company moves to the Cloud, its suppliers will also have to because they will now need to log into their account in the Cloud so that they can access the required data. If people continue to exchange files as they do now, then moving to the Cloud becomes a pointless exercise.

Personally, I am of the opinion that large companies will be more open to moving to the Cloud than small companies and individuals. But then you only need the large companies to make the move and the rest will automatically follow, whether they want to or not. I know this sounds offensive. But this is just how it has been for years and I see no reason why it will not continue to be.

Another thing that people are all bent about is having their CAD vendor host their data. Henrik Vallgren commented, “CAD in the cloud is taking trust into the extreme. Who would trust a company that encrypts the data files it creates on a PC?”

Actually, the question itself is its own answer. I mean, in spite of knowing that CAD vendors encrypt data people still use their software. My point is that users already trust their CAD vendors more than they realize. There is an individual called Paul Waddington who has been screaming himself hoarse for years about Autodesk not disclosing exactly what data their software transmits to their servers via the Customer Involvement Program (CIP). And yet, people sign up for the CIP program. How do you know that Autodesk does not send your drawings across to their servers when you connect to their serves to update the software, offer CIP data and a multitude of other reasons? You don’t. You simply trust Autodesk. Sames goes for every other vendor.

I have a bank locker where I keep some important documents and valuables. The locker works with two keys, one is with me and the other is with a bank official. You need both keys to open the locker. How do I know that the bank official did not make a duplicate of my key. After all it was he who gave it to me in the first place. I don’t. I trust the bank to hire trustworthy employees. My bank does not make money by robbing and selling personal belongings kept by its customers. It makes money because people trust it with their money and belongings. CAD vendors do not make money by stealing someone’s patented drawing or intellectual property. They make money by licensing software to people who trust them. If people don’t trust CAD vendors now then why the hell do they use their software? How does the Cloud make this so vastly different?

I develop CAD software. If I want I can make my software send me all kinds of data belonging to my customers. The moment someone runs my EXE on their computer and tells their firewall that its OK for that EXE to connect to the internet, I can do just about what I want on that computer, just like a virus. But I don’t because that is not my freaking business. And that is why people buy my software.

What the heck would Autodesk do with an architect’s drawing. Build a house for Carl Bass? Or sell it to another architect? Personally, I find this whole “I cannot trust my CAD vendor” argument rather stupid because you are already trusting your CAD vendor.

Stop for a while and think about it.