wmb@MITCH.ENG.SUN.COM (Mitch Bradley) (12/24/90)
Mitch Bradley: > >I have been publishing a stdio-like file interface for seven years... > >the result: all serious vendors have a file interface... Ray Duncan: > This is a serious distortion of the facts. LMI, for example, has been > selling file-system based Forths for much longer than this. Creative > Solutions, Micromotion, and other progressive companies also saw the > need for file-based Forth systems very early on and pursued this > aggressively. Mitch Bradley: I apologize for stating this unclearly. I did not mean to imply that my file system interface caused vendors to adopt a file interface. Indeed, I think just the opposite; it was market pressure and opportunity that led vendors to adopt file system interfaces. The point is not about "who had what when, and why". The point is that I have been working diligently for years to promote the idea of a portable file system interface, using avenues outside the context of a published standard, and have failed. Most (all?) serious vendors these days do indeed have quality support for text files, but there is little compatibility among the various vendors in this area. I cite this experience in support of my claim that standards are the best, and perhaps the only, way to add portable "improvements" to Forth. My view of the cited history: (The following is a statement of the what happened as I saw it. Other facts outside my "field of vision" will doubtless contribute to a more globally-accurate picture. This is only a "distortion" to the extent that the facts as seen by a single observer can be called such.) Creative Solutions did not "aggressively pursue" a text-file based system until some years later (based on my recall that the first version of Mac Forth, circa 1984, was block-based. The paper in question was published in 1983.) It is true that Master Forth adopted a file system interface at about the same time. At the 1983 FORML conference where I presented the paper, Martin Tracy showed me a pre-publication draft of Mastering Forth with a file system interface wordset described therein. As a result, I changed some of my word names to be compatible with what Martin published. I do not know about LMI. I do know that, 2 years later, I did some work on a product that was written in LMI Forth for the PC, and it could not compile from text files. I am willing to believe that the company may have been using an old version, or may have chosen not to purchase the version with the file interface. I do clearly remember that the programmer I was working with was under the impression that LMI Forth did not support text files. He may well have been wrong. Mitch Bradley, wmb@Eng.Sun.COM