Objects by Design Forums Pages (2): « 1 [2]
Show 15 posts from this thread on one page

Objects by Design Forums (http://forums.objectsbydesign.com/index.php)
- UML Questions (http://forums.objectsbydesign.com/forumdisplay.php?forumid=6)
-- Data modelling vs object oriented modelling (http://forums.objectsbydesign.com/showthread.php?threadid=467)


Posted by daybyter on 03-26-2003 04:26 PM:

quote:
Originally posted by Scott Ambler
[B]I think that there are several issues with the Rational proposal:
1. It's not complete. Mine isn't either, but it's certainly a lot more along the road.



Well the docs they posted are definitely not complete. I thought their tools implement the complete profile, but since I don't have them, I could never really check that.
One of my ideas was, that I should come up with a website, that shows the details of my implementation. Well the concept was, that we might try to set our own standard, if there is no definitive standard yet, But this was just, because the Rational papers are too incomplete to act as a manual on the profile.

quote:
2. It's vendor specific. I suspect that the primary reason why it isn't complete is that they don't cover the items they can't implement right now.


Well, I thought that it's more complete than the docs I've seen, as I said...
I've had a few problems with it, too. So I simply mailed Mr. Gornik and got the details, that the papers don't include.

quote:
Also, do you really want to rely on a competing vendor for the definition of what you're doing?


Not really, but I thought that Rational is the best candidate to come up with a de facto standard, that might be accepted by the OMG later. You have to consider, that I've seen your version just today and could only take a quick look, so I cannot really comment on it.

quote:
3. It doesn't look like they're going to submit to the OMG any time soon. A couple of years ago I was working with them to flesh it out and the effort sort of petered out.


That was something, that really puzzled me. I asked Mr. Gornik, but got no real reply. Well. the reply was, that he cannot talk about this issue at the moment.

quote:
4. If you read mine closely, you'll see that it's pretty much a superset of the Rational proposal (why reinvent the wheel).


It seems, there are some differences, like the steretypes for the classes (from what I know now)?
Is there any tool yet, that implements your profile? My idea was so far, that I'd like to come up with something very simple, but inexpensive. I'd like to combine it with a PHP module, I'm working on, so it could be used to model the database layer of website (but not exclusively, since you could also combine it with Java as an example).

Ciao,
Andreas


Posted by Scott Ambler on 03-26-2003 04:40 PM:

Well the docs they posted are definitely not complete. I thought their tools implement the complete profile, but since I don't have them, I could never really check that.

Scott: I haven't done an exact analysis, but I suspect there's a one to one mapping of the profile to what their tool supports.

One of my ideas was, that I should come up with a website, that shows the details of my implementation. Well the concept was, that we might try to set our own standard, if there is no definitive standard yet, But this was just, because the Rational papers are too incomplete to act as a manual on the profile.

Scott: I'm happy to work with others on the profile. Any feedback would be welcome. I definitely recognize that more work needs to be done on it.


Well, I thought that it's more complete than the docs I've seen, as I said...
I've had a few problems with it, too. So I simply mailed Mr. Gornik and got the details, that the papers don't include.

Scott: Take a look at my profile. Like I said, I'd be happy to work with you on it, share the information publicly (I try to act quickly on stuff like this, and usually do), and move forward on it.


-------------------------------------------------------------------------------
Also, do you really want to rely on a competing vendor for the definition of what you're doing?
--------------------------------------------------------------------------------



Not really, but I thought that Rational is the best candidate to come up with a de facto standard, that might be accepted by the OMG later. You have to consider, that I've seen your version just today and could only take a quick look, so I cannot really comment on it.

Scott: Check it out and let me know what you think.


quote:
--------------------------------------------------------------------------------
3. It doesn't look like they're going to submit to the OMG any time soon. A couple of years ago I was working with them to flesh it out and the effort sort of petered out.
--------------------------------------------------------------------------------
That was something, that really puzzled me. I asked Mr. Gornik, but got no real reply. Well. the reply was, that he cannot talk about this issue at the moment.


Scott: How about that. ;-)

Scott: To be fair to Rational, putting something through the OMG process is a horrendous and thankless task. Also, others have tried to do this and failed. From what I can tell in part due to a prejudice by OO folks against data issues and in part due to politics between CASE tool vendors. I personally wouldn't want to go through that effort myself, hence my grass roots effort.


--------------------------------------------------------------------------------
4. If you read mine closely, you'll see that it's pretty much a superset of the Rational proposal (why reinvent the wheel).
--------------------------------------------------------------------------------
It seems, there are some differences, like the steretypes for the classes (from what I know now)?

Scott: There definitely are differences, many of them extensions. I'm not constrained by what Rose can implement and I started with requirements (see the end of the essay) so may have had a better handle on what was needed.


Is there any tool yet, that implements your profile? My idea was so far, that I'd like to come up with something very simple, but inexpensive. I'd like to combine it with a PHP module, I'm working on, so it could be used to model the database layer of website (but not exclusively, since you could also combine it with Java as an example).

Scott: Not that I know of yet, although I have been talking with several vendors so I do expect something to come out eventually.


Ciao,
Andreas

__________________
Scott W. Ambler


Posted by daybyter on 03-27-2003 01:39 PM:

It becomes work to get the quotes right and since I guess noone here is interested in the discussion, we might continue the discussion via p-mail? mail@andreas-rueckert.de

quote:
Scott: I haven't done an exact analysis, but I suspect there's a one to one mapping of the profile to what their tool supports.


I could live with that just fine, if they'd post excactly, what their tool supports, so we could get some kind of interoperability...

--<snip>--

quote:
Scott: I'm happy to work with others on the profile. Any feedback would be welcome. I definitely recognize that more work needs to be done on it.


Ok, I'd like to focus on some core features, that would help to create a physical model of a DB, so my MySQL folks could generate code from it and I could RE their running DBs into diagrams. Most of the potential customers don't have any concept of UML or modelling in general, so I don't want to confuse them more as really necessary. OTOH, interoperability with future tools might be helpful. Which would require exact specifications of features like column storage (read: chapter 4 of your profile is way too short). I guess you want to store stuff like datatypes (of physical DB models) as tagged values, too? I just had a couple of minutes to read, but that's my best guess so far. If I'm correct, than the tagged values have to be defined exactly, so a XMI representation of a data model could be postprocessed with other tools. I mailed Davor Gornik about this problem and he mailed me some of his names, that we could use?


quote:
Scott: Take a look at my profile. Like I said, I'd be happy to work with you on it, share the information publicly (I try to act quickly on stuff like this, and usually do), and move forward on it.


As I mentioned, I'd definitely want to standard that is good enough for tool compatibility if both tools implement the entire profile. So it's basically about the names of stereotypes, tagged values etc.

Your concept of index classes will cause me some trouble, but maybe we can find a workaround or compromise (I thought the concept of a operation is not that bad, since a index might be computed).

--<snip>--

quote:
Scott: Check it out and let me know what you think.


Had a couple of minutes as I wrote, but I guess there's a lot more to come.

--<snip>--

quote:
Scott: To be fair to Rational, putting something through the OMG process is a horrendous and thankless task. Also, others have tried to do this and failed. From what I can tell in part due to a prejudice by OO folks against data issues and in part due to politics between CASE tool vendors. I personally wouldn't want to go through that effort myself, hence my grass roots effort.


Well, I'm not too happy with the OMG, too. They should provide at least a open forum, where one could report bugs in their current specs.

--<snip>--

quote:
Scott: There definitely are differences, many of them extensions. I'm not constrained by what Rose can implement and I started with requirements (see the end of the essay) so may have had a better handle on what was needed.


Ok, good news for me is, that I can ignore a good part of those specs, since MySQL doesn't support them or they aren't very useful in the context I'm focussed on. They might be added in later versions.

--<snip>--

quote:
Scott: Not that I know of yet, although I have been talking with several vendors so I do expect something to come out eventually.


Well, I'll just hope that they will be too expensive to pose a competition to my project.
I'm thinking about something really lowcost (maybe 100 bucks, or so), so tools like Together are not really a problem.

Ciao,
Andreas


Posted by umluser on 06-13-2003 06:59 PM:

Scott Ambler -- a junior member z1?? :)

I just had to laugh when Scott Ambler is on this forum as "junior member" )

It is an honour to have Scott Ambler here and read his views on this topic.


All times are GMT. The time now is 11:08 PM. Pages (2): « 1 [2]
Show 15 posts from this thread on one page

Powered by: vBulletin Version 2.3.5
Copyright © Jelsoft Enterprises Limited 2000 - 2018.
Copyright 1999-2005, Objects by Design, Inc.