Hi Otis

Just want to know what you think - is Objectspaces a threat and if it is - how serious? I know it is not out yet - only 2005.

Objectspaces is dead. The team is moved to the WinFS team, Objectspaces' API is scheduled to be merged with WinFS' API and will change severily.

WinFS will be released with Longhorn, and I estimate that will be mid 2006/end 2006.

It was never a threat really, only when it looked like MS would release a free mapper tool to create Objectspaces mappings and when objectspaces would be a key part in .NET. After they made a couple of mistakes (the API was horribly tied to Yukon) and after I did the math: 4 years with 4 people brought them a very badly designed API, I came to the conclusion: it was never meant to be.

Thomas Tomiczek of EntityBroker asked me earlier this year if I would join him in a lawsuit against MS because of illegal anti-competitive practises of MS. MS behavior related to Objectspaces was, at that time, indeed illegal, looking at EU anti-trust law, however it probably would take 6 or more years to get the case to an end, bringing a lot of bad publicity and what's worse: a very high lawyer bill to the table, so I did not join him (I did get him a decent lawyer though via my brother in law, who is a lawyer) (there was nothing to win: even if you win after 6 years, objectspaces (and LLBLGen Pro as well) will be old tools).

Thomas did proceed with the lawsuite but put it on hold now as Objectspaces is officially canned . I would lie if I'd say that I wasn't relieved when I read the first message about Objectspaces' cancellation, however as I said earlier: it wasn't a real threat anymore, but it would of course draw attention away from other tools towards MS' tool.

I had and have the feeling something else was going on behind the scenes for some time already as you have to be a very lame developer to produce the poop Objectspaces was in the march build after 4 years (!) with 4 people (!). So I have the feeling they worked on WinFS already for over a year and objectspaces was more of a petproject internally (the same as ASP.NET once was (ASP.NET 1.0 is written by just 3 people )

What's a bigger problem, and what does concern me, is the lack of clear focus at MS. Not only did they have to make cancellation statements in the case of Objectspaces, the merge with WinFS is also a weird one: it implies that Yukon will not be seen as a desktop database, because WinFS is the desktop database of choice. If that's not the case, why going through all the trouble of having a 'unified O/R layer in windows' as A. Conrad said yesterday (Andrew Conrad is the Objectspaces PM) on his blog: http://weblogs.asp.net/aconrad

This is just a hunch, but I won't be surprised if this lack of clear focus will make Yukon slip for another 6 months or Longhorn slip for another 6 months. WinFS looks very cool, however you have to be a real dump developer to call your universal datastore a 'Filesystem' (WinFS)... (or better: you must be working at marketing)

Also, MBF (Microsoft business framework) relies on Objectspaces as well, at least it did. As it is postponed till Longhorn as well, who knows what they will do with MBF as well...

mid/end 2006 is too far away to make clear plans for tools working with .NET, as the specs for .NET 2.0 aren't even finalized yet.

Thanks for the lengthy reply

It is good to know Objectspaces is officially canned - don't want to be in a situation where a cleint decides 2 or 3 years from now that they rather want an Objectspaces implementation just because in comes from M$ - The people funding projects (Accountants, CEO's, Directors...) usually want to stay away from non M$ technologies - Idiots - because they want to be backed by big corporations like M$.

Coming from a very MS-oriented company, I think LLBLGen is the first non-MS tool we have used for a very long time! The trouble is as a time-pressured small development team MS can make it very hard for you to look elsewhere....or rather they make it very applealing to use their tools and technologies.

I didn't know much about o/r mappers before llbl really. However, I thought that ObjectSpaces was non-MS tool, and MS had taken over a company to get hold of an o/r mapper rather than develop one - where the hell did I get that from?!

I used to read a lot of postings on asp.net/forums - in the architecture section. In there a few people were of the opinion "I am not going to use a O/R mapper because eventually MS will release one and will 'force' me to use that" - I think some developers are still of this opinion and are holding back from using an o/r mapper because they think ms are on the verge of sorting out all their problems with the magical objectspaces!