The ORM Foundation

Most people who use this site are working on proprietary systems that they can't share with other people, but I wonder about having an area where people can post their data models and get feedback from other users. NORMA doesn't currently support the use of external models, but maybe if we had a community library of common models then we could be ready for when that does happen. In my current consulting projects, there's no reason to keep the models themselves private. There's no "secret sauce" that a thousand sites haven't copied. The difference is that I'm using ORM instead of pre-packaged databases. In the long term, this could be a part of the forum that helps codify techniques for driving application development from conceptual models.

Hi Tyler,

That's a great idea!
What I can do is to set up a new section in the Library and a new section in the Forum.
What do you think these sections should be called and where do you think that they would best fit into the existing structure.
Ken

 

Perhaps something like outlined below? Existing sections are bold, new sections are underlined, comments on what the section is used for are italic.

• Forums
• Models
• In Progress "I'm modeling a situation, here's what I've come up with, please tear it to shreds and point out the weaknesses." Friendly discussion only. :)
• Requests These are common situations encountered in the field, somebody should put something together and fix it for all of us. Could have a thread for bounties if anyone knows of government grants or companies that need custom models.
• Reverse Engineering (non-proprietary only!) Taking open-source projects (osCommerce, DotNetNuke, Sugar CRM, etc.) and making ORM diagrams of the existing systems. Then cleaning things up, of course!
• Standards
• Models Discussions of making ORM models of standards. The end results would end up in Library -> Models -> Standards
• Library
• Models
• Snippets Small models that demonstrate specific concepts. US Addresses, Thai names, Australian beer sizes, etc. Alone, these models aren't terribly useful. They do serve to clarify ideas and build larger systems from.
• Systems Complete systems such as CMS, message boards, resource scheduling, etc. Larger than snippets, the resulting databases are complete and stand alone. Preferably models that have been field-tested and bullet-proofed.
• Standards Implementations of government specs, ISO standards, etc.

[EDIT: formatting]

OK, That's a good suggestion for consideration.

The other dimension that people may find of interest is to devise a way to classify the "situation models" and the related Forum discussions by "recognisable subject area".

For example: It seems to me that all organisations have a need to use "address models" for data on employees, customers and clients. However, there may by differences in the requirements between industry segments.  e.g. The organisations that issue passports may have different needs (such as a need to hold biometric records) that are not needed by a mail-order company. 

I like your "standards" idea. I tried this in the early 1990's when I served as Secretary of the Business Information Modeling groups for EDI in ANSI and UN/EDIFACT. At the time all we had was FORML and InfoModeler 1.0.  ORM has evolved a lot since then so we should be able to be more effective.

What are your thoughts on the best way to interest people in participating in the model groups?

Do you have in mind "NORMA only" models or or should we include VEA as well?

Ken

Absolutely, the "situation models" is a great idea. It might be nice to see a hierarchical tree of situations like this:

• Address
• Europe
• Germany
• France
• Netherlands
• UK
• America
• Canada
• Mexico
• U.S.
• Taxonomy
• Horticulture
• Medicine

The alternative is the "tag cloud" where people would add "Address, Netherlands, Model Request" to the thread. Between the two systems, I like the visual appeal of the hierarchy but the tag cloud would probably be easier to implement since there's already a Tags option for posts.

Tags may be more robust in the long term. Not everything is easily classifiable, and biometrics, for example, may pop up in more places than passports.

As for which models to post... I'm very NORMA-centric, but I see no reason to limit it to that. This is all about practical model-driven development, so if people are using VEA they should post VEA files. In a perfect world we'd have a converter running on the back end that would store the models, then spit out whatever file format people wanted. While we're dreaming, it'd be great to have that storage system generate screen shots of the models for people to browse through without having to download them. Realistically, I'd be happy just for a place to get feedback on my rusty modeling skills.

As for getting people interested in participating, the "if you build it, they will come" philosophy is kind of what I'm hoping for. From what I can tell there are no existing ORM repositories where people are sharing and collaborating on models. Of course, my research is limited by the fact that searching for "Object Role Modeling" turns up pages of academic papers, and any search for ORM is going to pull up what I call "The Dark ORM".

I don't know much about the international community other than this forum, but there are plenty of Neumont alumni who are out in the field and their colleagues say, "It's a pretty picture, but what can you do with it?" If they want to use it, they have to model nearly everything from scratch. Eventually they give up because it's easier to add a few tables, use The Dark ORM to reflect on it, and hit the next deadline.

Another "perfect world" thought: It would be nice if there were some kind of version control for the models, especially for the ones that are under active development. Being able to pick a version to modify, add a branch for your recommendations, etc., would be very helpful from a collaboration standpoint. Kind of like Sourceforge for models, but with less overhead for creating a new "project".

What I had in mind was what you have called "taxonomy".
My feeling is that this is a multinational dimension.
For example, a model about say "manufacturing shop floor management" would be relevant to any factory(sorry "plant") wherever it is located in the world.

More later...

Ken

Yep! Another good idea. 
The only way I can think of to do it at present is to upload files with suffixes such as _01, _02 and so on.

Over the last three years, I have researched several open source packages for version control.
I have not yet settled on one. If you know any that use ASP.NET /SQL Server - let me know.
The alternatives are to have a separate web site for this or to add a new page to this site. Both alternatives require quite a lot of work.

Ken

 

Like I said, that would be the dream system. Maybe that can be one of the first requested models in the message board-- a model to handle a bulletin board system where users can upload version-controlled files! In the meantime there's a lot of ground to be gained just by having a library and a discussion thread.

OK - I have set up an "open source models"  section in the Forum and in the Library.

I have to specify file types, so please let me know what file types would be good.

So far I have enabled .doc, .pdf and .orm. 

Ken 

Many thanks! I'll post a model or two tonight to get the ball rolling.

Good!

If you are interested, I can give you a blog page on this site so that you can add commentary to the OSM forum and OSM library.
(Oh dear YAA (= Yet Another Acronym)

Let me know if you want to do this and I'll set one up for you on a "let's see how it goes" basis.

Ken

Oh, but didn't you know software development was rife with TLAs (Three-Letter Acronyms)? The blog could be interesting. My style is a bit rambling, so I may end up scaring off more people than it attracts. On the other hand, it would tend to make my posts shorter if I was putting the domain and development commentary somewhere else. I'm willing to give it a try if you think it would be worth an experiment.

The blog setup worked fine when I ran a test on my local development installation.
It seems to make sense to setup your live blog after you have posted your first model to the Library.

So I'll wait until then - since it seems best to wait until you have something to blog about!

FYI: I'm planning to post a model that I prepared last January.
However, it will be a day or two before I have prepared a properly annotated Word document... 

Ken

Ken Evans:

I have set up an "open source models"  section in the Forum and in the Library

 

I note you've added the new Library section under Tools -> Norma OSM... which is odd, because the first thing I was planning to do with posted models was convert them for CQL and repost then that way. I have a number of models posted in ORM, PNG and CQL on my website http://dataconstellation.com, but I see no good reason to repost them here.

Ken Evans:

I have to specify file types, so please let me know what file types would be good.

Well... .cql would be good 

Also... I don't see any new forum section yet.

WRT version control, http://github.com implements a good model; each person can "fork" anyone else's project and publish their own version. For file types where it works (i.e. not ORM or other XML formats), merge changes (and even track changes) from any number of other people's forks of the project. In fact there is no true "master" version; each fork has an independent life. Obviously, in such a model, mere version numbering is impossible.