I think another fun way to determine how many votes people on X3J3 get is to come up with a test of ones' _current_ perceptions of Fortran and it's place in the industry. It should be mostly multiple-choice/true-false but perhaps some essay questions as well. For example (with answers at the bottom of the post, but not upside-down): -------- Question 1: The forthcoming Fortran standard [now F90] is important because: a) It will make it easier for people to write useful Fortran programs b) It will give compiler vendors more work to do and products to sell c) It will make it easier for people to teach others how to write Fortran d) It will serve as the ideal application lanuguage for Windows 3.0 Question 2: The Fortran 77 standard was important because: a) It was the first time anyone ever wrote down what Fortran was supposed to be used for b) Its deadpan writing style was the perfect antidote to the '70s disco craze c) Nobody thought it was possible d) It gave compiler vendors more work to do and products to sell e) It made it easier for people to write useful Fortran programs Question 3: The arithmetic-IF (three-way) statement is in the Fortran language because: a) IF statements having higher prime numbers of branches (five, seven, and so on) were found hard to implement on binary computers b) It used to be the only way for people to write useful Fortran programs c) It more closely models human expectations than the logical-IF statement, since at most intersections, a driver has three choices as to which direction to take d) It makes writing applications for Windows 3.0 easier Question 4: NAMELIST was added to the forthcoming Fortran standard because: a) There was no other way to make people use it b) There was no other way to stop people from using it c) It makes writing device drivers for UNIX easier d) People claimed it would make it easier for them to write useful Fortran programs e) Visual BASIC has it Question 5: Recent and future Fortran standards disallow multiple-dummy and/or dummy/common aliasing of any variable when the called procedure modifies any variable involved in the aliasing because: a) It will give compiler vendors less work to do and faster products to sell b) There should really be only one way to skin a cat c) Fortran programming is for people who have nothing else to do but remember obscure rules like this d) It makes it easier for people to write useful Fortran programs d) I don't know what a dummy is Question 6: Despite the fact that almost all Fortran implementations are on machines that use binary arithmetic, numerical constants in Fortran are expressed in decimal (base 10) notation because: a) It makes it easier for people to write useful Fortran programs b) There's only so much a computer should know c) Lots of Fortran programmers are used to COBOL PIC(999) stuff d) Keypunch machines are notoriously difficult when it comes to punching hexadecimal (base 16) numbers e) It will give anal-retentive mathematicians more explaining to do about how you never quite get what you want with floating-point, and that seems to keep them happy Question 7: The name FORTRAN itself means: a) FORmula TRANslation b) FORTy RANdom features in one language c) FORget your computer-science TRAiNing d) FOR The Right ANswers e) Fortran Only Resembles Text Remotely At Night f) Nothing, it is one of those made-up marketing names like MUMPS -------- Here are my proposed "points" for the above questions, and the reasons why: (Following the form-feed, of course.) Question 1: a) 10 points -- You can't argue with this and be useful to X3J3 b) 10 points -- You can't argue with this and be useful to X3J3 c) 9 points -- Point off for thinking teaching is as important as selling and using, since there's little money in teaching d) 0 points -- Try again when we start the standardization process for Visual FORTRAN Question 2: a) 2 points -- Nice fantasy b) 3 points -- True, but we fear your antidote to New Age music c) 4 points -- Doing the impossible gets boring after a while on X3J3 d) 10 points -- Cynicism important for X3J3 reps e) 10 points -- Cynicism important for X3J3 reps Question 3: a) 1 point -- For at least keeping up with comp.arch b) 10 points -- Your grasp of history is impeccable c) 5 points -- Always good to model human situations, but think: are such intersections ideal, or perhaps modeled on Fortran? d) 0 points -- Not true, since mouse up/down/drag status is not available as an integer value Question 4: a) 0 points -- Oh come on! b) 1 point -- Right idea, wrong application c) 0 points -- This is never a reason for a Fortran feature, and wrong too d) 10 points -- Truly, there seems to be no other answer e) 5 points -- Excellent abstract thinking, points off for being mistaken Question 5: a) 10 points -- A longer answer is possible, but a waste of time b) 9 points -- Not quite as eloquently expressed as a) c) 8 points -- Mostly true, but if we can get others to try it as well, compiler vendors can sell more compilers d) 4 points -- Only if you replace "easier" and "useful" with "possible" and "fast" e) 10 points -- You are likely to be excellent in X3J3 diplomacy Question 6: a) 10 points -- Strangely, most people still think in base 0xA b) 4 points -- True in sentiment, false in implementation, decimal gives the compiler _more_ information than it needs, sometimes c) 2 points -- It doesn't help them any, actually; COBOL programmers don't even understand what is meant by "Division" in Fortran d) 7 points -- Modern keypunch machines have their own hex entry pads e) 10 points -- The scary thing is, they're right Question 7: a) 10 points b) 9 points -- One point off for minor historical inaccuracy c) 4 points -- Cynicism not _that_ important to X3J3 d) 5 points -- Wrong, but simple-minded answer suggests malleable X3J3 voter e) 1 point -- X3J3 hardly needs more recursive-acronym-loving FSF weenies f) 0 points -- Not even close -------- NOTE: This sample test is protected by the GNU Public License. You may redistribute it only if you include with the distribution all the answers and free copies of the Fortran 66, Fortran 77, and Fortran 90 standards, plus MIL-STD 1753, for the recipients' background reading. tq vm, (burley) -- James Craig Burley, Software Craftsperson burley@gnu.ai.mit.edu Free Software Foundation 545 Tech Square Cambridge, MA 02139