What ever happened to ObjectSpaces?
Or more specifically how is this next version different from ObjectSpaces?
Why didn't ObjectSpaces ever ship? I remember first playing around with ObjectSpaces almost 4 years ago! We were continually promised that the ship date was "just around the corner". I think the last I heard was that it was going to be merged into the now-dead (castrated?) WinFS product.
What sort of lessons have you learned over the years from the failure of ObjectSpaces and the maturing OR/M market in general?
It seems like many vendors (Frans and his LblGen Pro product to mention one) and open source alternatives (Hibernate and it's red-headed step sister nHibernate) are already years ahead.
What I've seen so far seems encouraging but honestly not anymore mature than ObjectSpaces was at it's prime.
Ah yes... I remember Luca!
I wonder what ever happened to him?
He seemed like a great guy... very enthused about ObjectSpaces.
As far as I know, he's on the Linq for Sql team.
I believe it was Luca at PDC05 hosting the "Using the .NET Language Integrated Query Framework with Relational Data" session. In that sessioin he briefly mentiones ObjectSpaces and what specific ObjectSpaces learnings made it into LINQ to SQL.
Thanks for the responses guys!
I guess I'm just a bit confused as to where I should direct my attention.
Should I stick with something like NEO or LblGen or attach my barnacles to this new ship?
I guess it depends what I'm working on. If it's something that needs to get done now, the tried and tested options are probably the way to go. If it's something I'm just playing around with, maybe ADO.NET vNext is the way to go.
It's interesting to see you guys hanging around here. On one hand I'm sure you're interested in seeing what "the competition" is up to, on the other hand it is your competition and they're giving away their product for free (I know some ORMs are free already).
"I guess I'm just a bit confused as to where I should direct my attention."
If you are open to direct your attention to other alternatives that are shipping today, please check out NJDX, a .NET OR-Mapper based on the following KISS (Keep It Simple and Straightforward) principles:
http://www.softwaretree.com/products/njdx/whitepaper/KISSPrinciples.pdf
NJDX is simple, fast, flexible, non-intrusive, dynamic, lightweight, database-agnostic, and tightly integrated with Visual Studio. Avoiding code generation hassles, NJDX is easy-to-use. Free 30-day evaluation version is available from http://www.softwaretree.com.
-- Damodar Periwal
 Eric Kinateder wrote: |
I guess I'm just a bit confused as to where I should direct my attention.Should I stick with something like NEO or LblGen or attach my barnacles to this new ship?
|
|
You should use what gets the job done you need to do now. You can't use early betas of MS stuff in production now, and what's avaliable now will probably change before it's released. This thus means that if you're working on a project now, you've to get things done, today. So you should focus on tools which get that work done today.
I guess it depends what I'm working on. If it's something that needs to get done now, the tried and tested options are probably the way to go. If it's something I'm just playing around with, maybe ADO.NET vNext is the way to go.
|
|
New stuff gets a lot of attention in blogs/media etc. but this gives the false assumption that the majority is using this new stuff. That's not the case. Adoption of new products/tools is always done through phases, where a small group adopts it right away and if it's succesful, a larger group will follow etc. Most products never make it to the second phase.
So playing around because it's cool new tech, sure why not. Using it because you've to get things done... I'd opt for proven technology in this case.
It's interesting to see you guys hanging around here. On one hand I'm sure you're interested in seeing what "the competition" is up to, on the other hand it is your competition and they're giving away their product for free (I know some ORMs are free already).
|
|
The main problem with selling O/R mappers today is that you've to sell the O/R mapping idea as well. Once you've convinced the person O/R mapping might be a good solution for doing data-access, you then have to convince the person your toolkit is the best. With MS giving an O/R mapper which is OK but lacks some features here and there (it's a v1 toolkit, so of course they don't include everything) away for free, it will do a lot of evangelization of O/R mapping in the .NET world. You would be surprised how small the group of developers actually is who have actually heard of O/R mapping.
As long as we're staying ahead or provide additional value, it's ok.
You make some really good points.
I also see that vNext does not in fact mean v3.0, but some later release... pushing this out until 2007 or more likely 2008 or later.
Or... there is a slight possibility of never! :)
If I remember it right Luca was in a panel discussion at the last PDC.
All the guys in the panel were asked to told us what their wishes about Linq was.
Last in the row it was Lucas time to answer. His answer was "My wish is to finally ship something..."
A lot of laughs followed that.
It was very nice to listen to him, but I was a bit confused when I couln't realize why he talked about farms.
After I while I understood his english(with a spanish accent) and understood that he talked about forms...
Regards
Mikael Egnr
Luca's accent is not Spanish, but Italian :)
