Comments on: What is the future of search for PLM?

By: Debarshi Goswami

Debarshi Goswami — Thu, 11 Jun 2009 10:36:26 +0000

Hi Andrew, sorry for this late reply. I have tried out a product called 3DSearchIT. I like some features of this software. This is a shape based search engine granted US patent for technologies used in it. It’s search results are pretty good & rich. You can search attribute or a shape, you can also generate reports based on some particular search or all your database parts. More than that you can form groups of similar parts or duplicate parts & name them as per your requirement which may help you to manage huge number of parts very easily. Search techniques are awesome & results are very effective. Also you can view the difference between two parts in a separate viewer highlighting zones where they are different. This product has great potential & keeps updated all time. Parts of long range of CAD products such as Proe, STL, Solid Works, Inventor are supported. This is a product from Geometric Ltd, an Indian company works on CAD/CAM/PLM software. For you convenience I am providing you the link of this product.

http://3dsearchit.geometricglobal.com/

Hello Debarshi, Thanks for your comment! will follow links. -Best, Oleg

By: olegshilovitsky

olegshilovitsky — Wed, 29 Apr 2009 02:28:38 +0000

Andy, thanks for your insight and observations. I agree with your view on criteria->search->result->repeat. This is something different from web searches that focus on “result on the first page”. But, problem I see is that information is too heterogeneous. I’m not sure one solution is possible here, therefore I tried to analyze trends. For the moment, enterprise search looks like very unsuccessful. Most of them are focusing on content/document search in enterprise. 3D solutions are more mature, since they are based on previous CAD experience. -Oleg.

By: Andrew Sherlock

Andrew Sherlock — Tue, 28 Apr 2009 22:09:41 +0000

Oleg, thanks for the PartBrowser mention!

My thinking on this is that there are 2 aspects to what search for PLM should look like in the future:

a) the criteria by which you search (text, attributes, position, shape etc.).
b) the way those results are presented and visualized so that results can be assessed to allow for further navigation, filtering or comparison.

Search for PLM will have some commonality with more general enterprise search, but will also need some specific technology:

a) Criteria. Clearly, search of unstructured and semi-structured data that enterprise search copes with will be important. However, we also generate very rich, well defined models of the geometry of our products – clearly much richer than a textual description – and this information is hidden to traditional search. I think there are numerous circumstances where it should be searchable (‘have we made anything like this before?’) . But exactly how shape search and textual and attribute searches can be combined usefully, I think, is an open question.

b) Result visualization. The ways in which enterprise search products summarize and present results are optimized for textual data. I think there’s an opportunity to greatly improve on this for PLM data. Lists of part numbers, filenames or attribute values and perhaps dodgy thumbnail images aren’t a great way to present results from PLM search. Effective use of 3D visualization can be very beneficial.

Finally, I find it all gets rather confused when the term ’3D search’ is used. Does ’3D’ refer to the search criteria (ie 3D shape or 3D position you refered to)? or to the presentation of those results (i.e. 3DLive, PartBrowser or something like Cooliris)? Maybe we need a better way of talking about all this.

Andy