dph@lanl.gov (David Huelsbeck) (07/19/88)
A few weeks ago I posted a request for information about CASE tools and other program development environment thingies. Responses were heavily weighted on the side of "let me know what you find out". So here is what I found out. It was suggested that I attend the 3rd Symposium on Software Development Environments (SIGSOFT/SIGPLAN), to be held in Boston this November. I've made a request for registration materials and I hope to be there. A fairly broad survey of commercial tools was disappointing. It seemed that about 75+% of the commercial tools were those cute little MacDraw(tm?) type things for design automation. I guess that's OK if that's what you're into but I have serious doubts about their usefulness to us. (For some interesting comments on the usefulness of such tools see: _Automating_Software:_proceed_with_caution_, IEEE Spectrum, July '88) We did find a few commercial tools that looked interesting. These all fell into what the IEEE article referred to as "Implementation" or "Support" phase tools. The one draw back to these was that for the most part they supported languages we are not especially concerned with. (i.e. languages other than Fortran) We did find one commercially available set of tools for Fortran that we were genuinely interested in. It turned out that someone else here had already made plans to buy this package. So I guess we'll be getting a first hand look at it. In retrospect, none of the above ought to be a big revelation. I imagine that plenty of you are reading this and saying "Well, what'd they expect anyway?" I can only answer that by saying that I didn't really know what to expect. That's exactly why we did this survey. However, some good did come of it. I went ahead and did some more book work to augment this survey. I ran across an NBS Special Publication titled Software Development Tools. This was an '81 or '82 publication I think. (Sorry not to have the exact reference.) This included quite a few references to "public domain", Fortran oriented tools. We're now looking into the possibility of modifying one of these to meet our own specialized needs. (i.e. non-standard Fortran) This is something I think vendors should take note of. Outside of the Unix/C and Ada camps how many people actually adhere to standards strictly enough to make using a tool directed toward some ANSI language feasible? It may be a symptom of my both limited and unusual experience in this area but it seems to me that most vendors offer extensions and most users use them. It strikes me that a vendor that offered a set of implementation/support phase tools directed at a particular brand of Fortran or whatever could make a killing. I would think that as long as the user interface was slick, the tool did not constrain the users too much (i.e. would work on existing codes, it didn't require the use of some special CASE methodology), and it provided a few basic analysis features a market would be almost guaranteed. Are there people providing such flavor specific tools? Is it just that our particular flavor is so uncommon that no one provides tools for it? Are there any tools vendors out there that would care to comment? Thanks to those who responded. I hope this is of interest to some of you. David Huelsbeck Los Alamos National Laboratory dph@lanl.gov