The ORM Foundation

Sorry if this is addressed somewhere deep in the archives or something, but I'm wondering why not a standalone app that could be used with VS, with Delphi, with Net Beans or anything else involving a database? It seems being an addin to VS is overly limiting.

That's a great idea.

How much development effort do you think would be needed?
And of course where would the funding come from?

To quote from the movie The Right Stuff : "No bucks. No Buck Rogers!  

Ken

Darryl,

It would be great to have NORMA running integrated in multiple platforms. The original decision to target Visual Studio was based on the skillset of the students coming into the project, and also the comparative capabilities of the modeling toolkits (EMF vs DSL). The winning DSL feature that was unavailable in EMF (at least in 2004 when the project was started, I'm making no claims about 2012) was the ability to do delete closures, which directly enabled incremental error checking on deletion without revalidation of the entire model.

Just because NORMA runs in VS does not mean that artifacts supporting other languages and platforms cannot be generated with NORMA. You would just need to run the generator in VS. A number of NORMA features are available programmatically without Visual Studio running (model loading, verbalization, etc), but the code generation system (Tools/ORMCustomTool in the code base) is heavily tied into the Visual Studio custom tool architecture and would require significant development to run externally.

Most of my behind the scenes work has been on formal ORM-based derivations and browser-based model display and editing (you can see the display side of this following links at http://www.ORMSolutions.com). The framework behind this is declarative and designed around fact-based principles and is a much better match with ORM and declarative logic programming than either DSL or EMF. The goal is to formally model all of the NORMA functionality (ORM validation rules, relational model generation, etc) in .ORM models and then generate the structure and rules in both DSL and the web frameworks. At this point retargeting the generators to other frameworks (EMF, for example) would provide all of the non-UI portions of the tool. Of course, this is still a lot of work, but it will be significantly less work than attempting to rewrite several hundred custom rules from scratch.

-Matt

Ken:

     Okay, funding would be a big issue...unless lightning strikes and I win the lottery.......

Matt:

     That makes sense too, didn't know you were taking advantages of hooks in VS to make the work easier. Of course, my only worry there is when the day comes that MS deprecates VS in favor of the latest and greatest (i.e. Eureka...a much improved LightSwitch) and those hooks go away, or something.

Thanks.

There's also the APRIMO product I'm developing - not yet released - which is a web-based ORM2 modeling tool with real-time multi-party collaboration and versioning.

Its backend is the open source CQL language and ActiveFacts toolkit, so you can easily write your own SQL- or code-generators, or use the existing ones. See http://dataconstellation.com

Yes, funding is an issue. What wold you be prepared to pay in a monthly subscription for such a tool? This is a serious question - I'd like some real answers to be able to brief potential investors.

"web-based" kills it right there. Web goes down, I can keep on truckin...till the batteries go, anyway. Subscription? That's the second stake thru the heart. We've never gone either route, upgrading when the feature set made it worthwhile.

Well, that just says that you're not in the target market. That is, it doesn't kill the idea, it just excludes you from benefiting from it.

The modeling tool could easily be packed as a local-host served web app (i.e., works like a standalone app, but still in the browser). I just have no plan to do that yet. What would you pay for such a thing?

I lost the best decade of my life building and selling software tools for developers. It's a non-market; they have either no money or no willingness to spend it. The majority of the work in this project (well in excess of 90%) is for developers, and it's free open-source software. You just don't get the ORM2 diagramming and collaboration tools for free.

Hi Darryl,

Clearly this is a valid concern. I also don't like being tied to an Internet connection all the time. For example, I like having browser access to e-mail (so I can check up on things when I'm away from my own machine), but I would never trade the cached off-line access, load speed, search speed, backup and other features that I get with Outlook.

However, I don't think you should constrain the notion of web-based technologies to a live internet connection. Web servers are easily run on local machines now  (IISExpress, paster, etc), and fully functional html rendering controls are easily hosted in other applications (VS itself uses some html-based windows). Web-based also doesn't mean that the data needs to be stored somewhere in the cloud. For example, http://www.ormsolutions.com/tools/orm.aspx (with no query string) tries to open a local file (assuming an HTML5-compliant browser that supports opening local files). If you have a local web server running, you can grab the local source for this page, save it as an .htm file, grab the three referenced javascript files, and have a local orm file viewer running in your browser with no internet connection in a matter of minutes.

A big part of the issue with modeling is that many of the participants in the modeling process--both modeling experts and domain experts--are not developers, so requiring a heavy-weight development tool to create, view, discuss, and share models with all participants in the process can be a major stumbling block. Microsoft effectively killed the Visio-based ORM tools by packaging them in high-end, developer-targeted Visual Studio box. Simply put, we need to make it easier to use ORM models across a wider audience.

Of course, the broader availability of ORM is also the gist of your original query. I fully agree that we need broader access, and that developers and development tools are a big part of it. However, I also think that focusing too much on the back end technologies obscures the vision of where we want to be. The goal is to be able to design, implement, and execute at the conceptual level, with the backend becoming less and less of a concern over time. I think web-based technologies offer ways to approach this problem in a way that local apps and development tools can't touch. I'm not sure I would have said that five years ago, but the world is changing both in terms of connectivity and in the client-side browser capabilities.

-Matt

Matt,

Apparently I haven't been keeping up! I had no idea you were working on a web viewer. I've been thinking of diving in to contribute some UI nice-to-have things. Would it be better to contribute to the NORMA project, or are you more actively developing in the web version? I'm thinking NORMA first since I use it regularly, but a free web-based tool would be huge for popularizing ORM.

Clifford,

Sorry to be not in your target market either. I understand your pain, though! Recently I won some Scrum Master training from a local professional continuing-ed company. It got me thinking... ORM doesn't have anything resembling a certification process. It's premature, obviously, since we're a small minority of the modeling community. By popularizing the language, guys like us would be in a pretty good position to turn into modeling coaches!

Continuing that thought, I came across this today that's kind of along the lines of a mini training / indoctrination course for their particular tool. I don't know that the NORMA forum is necessarily the right place for this sales pitch, but some less formal YouTube introductory courses could be a step toward broadening our "market".