François Van Emelen writes: >Lack or absence of comments? >Here is one on Roy Wood's 'Very FEXing' (QLTODAY Vol 10 Issue 3). >He is so right in his article. It would have been so easy to to >avoid the problems with 'FEX'; a quick look at the 'keywords list' >(Dilwyn's and the official SMSQ website) would have told the >author that 'FEX' was already an existing keyword in one of our >most used utilities (FileInfo2).
Its all my fault! However, there are mitigating circumstances. First of all I had written a much simplified version of the FEW, FET and FEX keywords, which Id used for years before I ever heard of any other usage. Secondly, as Thierry wrote in the FI2 manual, the FEX keyword had been deprecated, that means that future versions of FI2 could well be produced without it. Before I decided to keep the FEX name in SMSQ/E, I performed a scan of all the software in my archives, commercial and PD. None used the keyword FEX (I obviously missed Norback). It would have been easy to avoid even the smallest risk of a conflict by choosing a different keyword, or set of keywords. But, considering the quantity of keywords built into Sbasic, I wanted to base the nomenclature on logic and precedent, as indeed Id done in my original toolkit. TT started the convention, for example by shortening the OPEN keywords to FOP_XXX function names. EXEC and EXEC_W had become EX and EW. It seemed logical to use the abbreviations FEX and FEW (and FET and FEP for those who know their TK2 and Qpac2). Wolfgang dutifully complained and we finally compromised by him supplying an extra handle to the function: EXF. I also produced a patcher that would patch any program using the original keyword FEX to something else. All of this was described in the documentation (except for some reason the patcher never got distributed). >My updated 'keywords list' have more than 3.500 entries now, but I >don't yhink it is still worth while publishing. > >He is so right being picky about the way 'SMSQ is updated. I agree with the sentiment about being "picky" about the way SMSQ/E is updated. I think it would be nice to keep the core vocabulary of Sbasic as "clean" as possible, avoiding unwieldly long names (as found in the Turbo toolkit, for example) or some of the very cryptic names found in some other toolkits, which is the reason I did it. Im sorry for the one little reported inconvenience caused and hope this has now been sorted. Per _______________________________________________ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm