Academic Facets: News and Views from Spatial's University Community

[#1]

The highlight of my week was a meeting of the UK’s Geometric Modelling Society (GMS) at BathUniversity, but sadly  I had to leave before Professor Chris McMahon inaugural lecture in the evening. Entitled “From cathedrals to cars and planes: representations in engineering design”, I was disappointed to miss it, but the title and some of the presentations at the GMS meeting provided food for thought during my journey home.
 

Digital preservation is an emerging topic in academic CAD/CAM research. Although it takes many forms it is all motivated by essentially the same problem, namely:  “when all mankind’s design data is digital how do you preserve it for future generations?”
 

It is easy, at first glance, to dismiss the subject as not being much of an issue, after all I can open the oldest ACIS sat file on my hard disk (1995) without a problem. But that is only 13 years old, the question entertaining University researchers is: ‘could you read that file in 30 or a 1000 years time?’
 

I think these are two different problems, progress over a thirty year time scale can almost be imagined (eg. perhaps subdivision surfaces will replace B-splines, perhaps facetted meshes will become the de-facto representation for CAM) whatever happens we are talking about a time span less than a professional engineer’s working life.
 

Not surprisingly there are already people wrestling with issues, and implementing solutions, for thirty year CAD archives. Last year, for example, at a workshop on “Long Term Knowledge Retention” (LTKR), I heard several speakers associated with various defence, aerospace and space industries talk about projects whose engineering data needs to stay “live” for many decades.
 

So what’s the strategy for today’s 30 year CAD archives? One clear trend was the adoption of STEP as an archive format. I don’t know if ISO anticipated this application of the standard but the public specification, and wide spread use, make it an obvious choice for any digital time capsules.
 

However while no one ever got fired for adopting STEP as an archive format there are other views. For example if there is a paradigm shift in mechanical CAD (say some disruptive modelling technology arrives from Gaming or Rapid Prototyping) and the B-rep does not survive as the central format, then what? Will the complexity of STEP files still be supported?

One insurance policy against this sort of eventuality is to use a less sophisticated representation like STL (the assumption being that even without any documentation many people could extract a basic model).  This sort of reasoning has led  CAD archiving researchers to become increasingly interested in so called “lightweight” representations like X3D, JT and Adobe 3D. Some of this work was presented at the GMS meeting by investigators from the UK’s KIM project (aka “Immortal Information).
 

If, they reason, use of these formats becomes widespread, then their long term readability becomes assured and although only geometrically approximate (compared to a B-rep) they could form a useful back-up to the detail of the STEP files.

The far future, is of course the most interesting and difficult. Given an “Ozymandias* scenario” (say CAD data for an underground nuclear waste repository being accessed after a thousand years) what can we do today, in anticipation of a time when perhaps ISO standards have disappeared and the computing technology is unimaginable?

Amazingly people are already employed to worry about  “1000 year issues”, for example at the same LTKR workshop (a rather impressive lady) from the US “Library of Congress” explained how her remit involved ensuring access to Catia v4 files (from Los Alamos) until “……the end of the Republic”! So what can be done today to facilitate 1000 year retrieval? The best notion I’ve heard is based on that exemplar of “extreme LTKR” displayed in the BritishMuseum…………the Rosetta Stone. The Rosetta Stone is an Ancient Egyptian artefact covered in carved text that repeats the same passage in three languages (Greek, Demotic and Hieroglyphic). Because 18th century scholars knew Greek and Demotic they were able, after a few years, to translate the hieroglyphics by comparing the known and unknown data.

I think the Rosetta stone tells us that rather than adopting a single CAD representation we should aim to build archives that hold many different representations of the same model (e.g. STEP and X3D and STL and …etc). If the future CAD users can understand one of these representations, then they stand a better chance of decoding the other ones.

I am not for a moment suggesting that this approach makes everything easy, there are still lots of problems: mapping non-geometric attributes from one representation to another (the persistent naming problem again) or cross format model validation. But if there were no problems there would be no research!  Anyway presumably 1000 years hence all the hard thinking about multiple CAD formats will be done by robots, so no need to worry.

 

*And on the pedestal these words appear:
"My name is Ozymandias, king of kings:
Look on my model formats, ye Mighty, and despair!"
Nothing beside remains: round the decay
Of that colossal wreck, boundless and bare,
The lone and level sands stretch far away