Is there a solution to the interoperability problem?

Over the past fews days I have been commenting on the Autodesk - ODA law suit. Although my observations have been strictly related to the law suit only, some may have got the impression that I was leaning towards one side. I want to keep the legal battle out of this one.

Interoperability is a big big problem for the CAD software industry. I will not waste my time and yours in describing the problem because, firstly, I don't know where to start, and secondly, we all are facing it, so we know all about it. The question is: Is there a solution?

In my view, yes. And you know what? Someone has already done it.

I usually don't go out of my way to praise a particular CAD software vendor, mainly because I hate being called somebody's pet, but also because I would like to maintain my neutrality. But I will make an exception in this case. I am talking about Robert McNeel & Accosiates and their CAD software Rhinoceros. McNeel has done it right and if all other CAD vendors followed suit, the interoperability problem would simply vanish.

When McNeel created Rhino, like any other vendor, they created a native file format for it and gave it an extension - 3DM (stands for 3D Model). But, unlike other vendors, they also founded the OpenNURBS Initiative, through which they offered free software libraries to read and write 3DM files. These libraries are available to everyone, even rivals. Their software license reads:

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software.

There are absolutely no strings attached. Strings such as "for evaluation/educational use only", or "not for use in a commercial environment". Commercial use is not only allowed, it is encouraged. There are no copyright or copyleft restrictions. Truly extraordinary.

By doing this McNeel has made it very easy for Rhinoceros to exchange data with other CAD systems. They even offer free plug-ins to other CAD systems, such as SolidWorks, to read and write 3DM files. No need to reverse engineer. No file corruption. No "trusted" and "non-trusted" nonsense.

And people like me have taken things even further. When we wrote TerrainCAD, our terrain modeling software, instead of creating a proprietary file format and adding to the mess, we simply used the 3DM file format as the native format for TerrainCAD.

Now imagine if all CAD software vendors did the same thing. End users would spend their time doing useful work as opposed to fixing badly imported drawings, or even worse, recreating a drawing from scratch. The interoperability problem would simply not not be there anymore.

If one vendor has done it and not gone out of business, there is no reason why others cannot do it.

ODA shoots itself in the foot

To justify the display of the now famous (or rather infamous) "TrustedDWG" alert message, Autodesk claims that DWG files created by ODA's DWGdirect libraries causes AutoCAD to become unstable. This is an excerpt from their initial complaint.

One problem with third parties' implementations of the DWG format is potential data corruption resulting in instability introduced into the files by competitive products. Autodesk's customer support personnel have logged numerous instances of Autodesk customers receiving DWG files from outsiders, then attempting to open the files with AutoCAD software, only to encounter serious errors.

In their reply, the ODA refuted by saying: "Autodesk has offerred no empirical evidence that its files are any more stable than others". Indeed, they didn't. And I wondered why, especially since any evidence, however small, could make their case rock solid. I investigated a little and came up with something truly interesting.

This is part of the DWGdirect 2.0.3 Release Notes (2.0.3 is the release in which the ODA added support for DWG 2007):

Bugs Fixed in 2.0.2
3565 DD writes DWG file which crashes AutoCAD.

3565 appears to the bug number and DD stands for DWGdirect. After digging more I found these bug fix entries for other DWGDirect releases:

1742 Linetypes with text are incorrectly saved to DWG.
2008 Acad crashes on a file saved by DD, containing an xref.
2211 AutoCad crashes on file saved by DDT - proxy issue.
2780 OdaMfcApp saves an invalid R12 file, which crashes AutoCAD.
2782 File saved to DXF 13 can't be loaded by AC2004.
2783 File saved to DXF13 can't be loaded by AC14.
2887 Acad crashes on attached file. DD audit/recover find no errors.
3150 Saved DXF file crashes AutoCAD.

The list goes on and on. Here we have the ODA (very honestly) admitting that their libraries cause AutoCAD to crash. They just shot themselves in the foot. Autodesk does not have to prove anything. The ODA did it for them.

This looks like clinching evidence to me. A smoking gun with lots of smoke. Its baffling why the Autodesk attorney's didnt use it in their initial complaint. Could it be that they didn't know about it? Seems hard to believe.

Autodesk VS ODA - Compromise or Knock-Out

Randall Newton's blow-by-blow coverage of the Autodesk-ODA legal fight at AECnews.com is truly commendable. The latest article gives us an idea of where things may be headed. The article ends with:

As ODA and Autodesk prepare for the January hearing, there may be some efforts to negotiate a truce. Edelson [the ODA attorney] had proposed to the court that a more neutral pop-up, one that identified a file as non-Autodesk-generated without dire warnings, could be an acceptable solution. “Let’s keep the lines open” he said to Jacobs [the Autodesk attorney] as they prepared to leave the courtroom.

Since the ODA went ahead and incorporated the TrustedDWG feature in their DWGdirect libraries, I assumed that they had already built a strong legal case in the event of a law suit. After reading the ODA reply to the initial complaint it does not appear to be the case. Their reply sounded more like their views described on their web site. Views which a majority of CAD software users may agree to, but which do not mean much in a court of law.

With the judge virtually throwing all their arguments out of the window, the ODA now finds itself with it's back against the ropes. Throwing in the towel at this stage may be the wise thing to do. In any case, it doesn't look like Autodesk is in the mood of showing any mercy. Maybe this was what they were waiting for all along.

After ODA, who's next on Autodesk's radar?

With the courts granting a temporary restraining order against the ODA, I wonder if this is the beginning or is there more to follow?

I have been keeping a close watch on Dr. DWG , a division of California Software Labs. They offer DWG read, write and view libraries similar to ODA's DWGdirect libraries. Probably not as exhaustive as the DWGdirect libraries but light weight and easy to use. These libraries support the DWG version upto 2006. Interestingly they have not yet released an update for DWG 2007. That may be either because they have not yet reverse engineered it or because of precisely the same reason which the ODA finds itself in trouble today.

Autodesk has tried to trademark the word "DWG". Evan Yares has an interesting article on that. In fact on Autodesk's "Legal Notices & Trademarks" page, DWG is listed along with quite interesting guidelines of usage, part of which I am listing here:

You should not adopt or use product, service, or company names that could cause confusion about affiliation with or endorsement by Autodesk. Do not, for example, cite to the DWG name or to any other Autodesk trademark at the beginning of, or otherwise as the most prominent part of, your product, service or company name.

You should maintain a visual distinction between your company and product name, on the one hand, and DWG or any other Autodesk trademark, on the other. For example, Acme Co.’s “ALPHABETA for DWG” is permissible.

You should not create or use any logos that include DWG or any other Autodesk trademark unless your use is pursuant to a license from Autodesk. For example, you should not create your own DWG compatibility logo without permission from Autodesk.

You should not register any Internet domain names incorporating DWG, unless such domain names are not misleading or confusing and unless there is clear and conspicuous trademark attribution to Autodesk on the homepage of the corresponding website.

In Dr. DWG's best interest, they had better not come up with DWG 2007 (with or without the RealDWG feature) untill this storm has blown over.

Did the Open Design Alliance walk into a trap?

With Autodesk suing the OpenDesign Alliance for Trademark Infringement (see full article here), I cannot help but think that the ODA walked right into a trap set by Autodesk.

It's no secret that Autodesk has continuously tried to make life increasingly difficult for the ODA. They have gone to lengths to make the DWG file format more difficult to reverse engineer, and for good reason (see my article 'Should Autodesk keep the DWG format a secret?'). Since this approach has not worked well, they have now gone the legal way. It appears that Autodesk left the ODA no option but to violate their trademark.

On May 12, 2006, the ODA had a press release titled 'Update on status of DWG 2007'. I found these parts particularly interesting:

"We have, at this time, reverse-engineered the high level format for DWG 2007. Based on the initial information we've received from the software engineers doing the work, it appears that the new file format is significantly scrambled at a high level. It seems that this might be at least partially a result of Autodesk including Reed-Solomon error correcting code (ECC) in the files. (ECC is a good thing, and if this is the only reason for the format being scrambled, we'd certainly have nothing to complain about.)"

"If this is the only reason for the format being scrambled". Hmmm.... They go on to explain about TrustedDWG, the thing which has come back to haunt them.

"The 2007 versions of AutoCAD-based products introduce a concept called 'Trusted DWG.' This is an update of the existing DWGCHECK functionality that has been in AutoCAD-based products for some time. Trusted DWG, from what we can tell, has little to do with DWG files actually being trustworthy—it is primarily a verification process that AutoCAD-based products do to check if a DWG file was last written by an Autodesk program. The mechanism used is an encrypted digital signature, placed in the DWG file (starting with the DWG 2004.) This digital signature includes a checksum (which verifies that the file hasn't been changed since it was written), and a code indicating which Autodesk product wrote the file. Up until this point, we've not provided an API to support reading or writing this digital signature, as it seemed to be fairly benign, and of little practical use. With the apparent inclusion of true ECC in the DWG format, it makes sense for us to start supporting this feature."

So the ODA never wanted to support the TrustedDWG feature because "it seemed to be fairly benign, and of little practical use". However, I believe the main reason for this decision was to avoid trademark infringement. It looks like they were left with no other choice.

There is absolutely nothing illegal about reverse engineering the DWG file format. As far as my understanding about trademarks goes, I can buy a bottle of Coke, analyse its contents in a lab and come up with the formula. Assuming patents are not involved here, I can then make a soft drink using that formula, sell it and make money. But there is no way I can label my product as "Coke" or something that is confusingly similar to "Coke". What Autodesk have done with TrustedDWG is exactly this. They have forced the ODA to support TrustedDWG, thereby making them violate their trademark. I may be wrong. Let me know.

Since the ODA went ahead and supported the TrustedDWG feature, I guess their lawyers have already built a case in defense. One thing is sure. The battle lines are drawn and this fight will be interesting.