Return to e6 CVS log

Up to [CSRG BSD Unix] / 43BSDReno / share / doc / usd / 26.eqn

Annotation of 43BSDReno/share/doc/usd/26.eqn/e6, revision 1.1

1.1     ! root        1: .\"    @(#)e6  6.1 (Berkeley) 5/22/86
        !             2: .\"
        !             3: .NH
        !             4: Experience
        !             5: .PP
        !             6: There are really three aspects of interest_how
        !             7: well
        !             8: .UC EQN
        !             9: sets mathematics,
        !            10: how well it satisfies its goal
        !            11: of being ``easy to use,''
        !            12: and how easy it was to build.
        !            13: .PP
        !            14: The first question is easily addressed.
        !            15: This entire paper
        !            16: has been set by the program.
        !            17: Readers can judge for themselves
        !            18: whether it is good enough for their purposes.
        !            19: One of our users commented that although the output
        !            20: is not as good as the best hand-set material,
        !            21: it is still
        !            22: better than average,
        !            23: and much better than
        !            24: the worst.
        !            25: In any case, who cares?
        !            26: Printed books cannot compete with the birds and flowers
        !            27: of illuminated manuscripts on esthetic grounds,
        !            28: either,
        !            29: but they have some clear economic advantages.
        !            30: .PP
        !            31: Some of the deficiencies in the output could
        !            32: be cleaned up with more work on our part.
        !            33: For example, we sometimes leave too much space between
        !            34: a roman letter and an italic one.
        !            35: If we were willing to keep track of the fonts
        !            36: involved,
        !            37: we could do this better more of the time.
        !            38: .PP
        !            39: Some other weaknesses are inherent in our output device.
        !            40: It is hard, for instance, to draw a line
        !            41: of an arbitrary length without getting
        !            42: a perceptible overstrike at one end.
        !            43: .PP
        !            44: As to ease of use,
        !            45: at the time of writing,
        !            46: the system has been used by two distinct groups.
        !            47: One user population consists of mathematicians,
        !            48: chemists, physicists, and computer scientists.
        !            49: Their typical reaction has been something like:
        !            50: .IP " (1)"
        !            51: It's easy to write, although I make the following mistakes...
        !            52: .IP " (2)"
        !            53: How do I do...?
        !            54: .IP " (3)"
        !            55: It botches the following things.... Why don't you fix them?
        !            56: .IP " (4)"
        !            57: You really need the following features...
        !            58: .sp 5p
        !            59: .PP
        !            60: The learning time is short.
        !            61: A few minutes gives the general flavor,
        !            62: and typing a page or two of a paper generally
        !            63: uncovers most of the misconceptions about how it works.
        !            64: .PP
        !            65: The second user group is much larger,
        !            66: the secretaries and mathematical typists
        !            67: who were the original target of the system.
        !            68: They tend to be enthusiastic converts.
        !            69: They find the language easy to learn
        !            70: (most are largely self-taught),
        !            71: and have little trouble producing the output they want.
        !            72: They are of course less critical of the esthetics of their output
        !            73: than users trained in mathematics.
        !            74: After a transition period, most find
        !            75: using a computer more interesting than
        !            76: a regular typewriter.
        !            77: .PP
        !            78: The main difficulty that users have seems to be remembering
        !            79: that a blank is a delimiter;
        !            80: even experienced users use blanks where they shouldn't and omit them
        !            81: when they are needed.
        !            82: A common instance is typing
        !            83: .P1
        !            84: f(x sub i)
        !            85: .P2
        !            86: which produces
        !            87: .EQ
        !            88: f(x sub i)
        !            89: .EN
        !            90: instead of
        !            91: .EQ
        !            92: f(x sub i )
        !            93: .EN
        !            94: Since the 
        !            95: .UC EQN
        !            96: language knows no mathematics, it cannot deduce that the
        !            97: right parenthesis is not part of the subscript.
        !            98: .PP
        !            99: The language is somewhat prolix, but this doesn't seem
        !           100: excessive considering how much is being done,
        !           101: and it is certainly more compact than the corresponding
        !           102: .UC TROFF
        !           103: commands.
        !           104: For example, here is the source for the continued fraction
        !           105: expression in Section 1 of this paper:
        !           106: .P1
        !           107: .ne 4
        !           108: .ce 0
        !           109:      a sub 0 + b sub 1 over
        !           110:        {a sub 1 + b sub 2 over
        !           111:          {a sub 2 + b sub 3 over
        !           112:            {a sub 3 + ... }}}
        !           113: .P2
        !           114: This is the input for the large integral of Section 1;
        !           115: notice the use of definitions:
        !           116: .P1
        !           117: .ce 0
        !           118: .ne 15
        !           119: .in 1
        !           120: define emx "{e sup mx}"
        !           121: define mab "{m sqrt ab}"
        !           122: define sa "{sqrt a}"
        !           123: define sb "{sqrt b}"
        !           124: int dx over {a emx - be sup -mx} ~=~
        !           125: left { lpile {
        !           126:      1 over {2 mab} ~log~ 
        !           127:            {sa emx - sb} over {sa emx + sb}
        !           128:    above
        !           129:      1 over mab ~ tanh sup -1 ( sa over sb emx ) 
        !           130:    above
        !           131:      -1 over mab ~ coth sup -1 ( sa over sb emx )
        !           132: }
        !           133: .in 0
        !           134: .P2
        !           135: .PP
        !           136: As to ease of construction,
        !           137: we have already
        !           138: mentioned that there are really only a few person-months
        !           139: invested.
        !           140: Much of this time has gone into two things_fine-tuning
        !           141: (what is the most esthetically pleasing space to use
        !           142: between the numerator and denominator of a fraction?),
        !           143: and changing things found deficient by our users
        !           144: (shouldn't a tilde be a delimiter?).
        !           145: .PP
        !           146: The program consists of a number of small,
        !           147: essentially unconnected modules for code generation,
        !           148: a simple lexical analyzer,
        !           149: a canned parser which we did not have to write,
        !           150: and some miscellany associated with input files
        !           151: and the macro facility.
        !           152: The program is now about 1600 lines of 
        !           153: .UC C
        !           154: [6], a high-level language reminiscent of
        !           155: .UC BCPL .
        !           156: About 20 percent of these lines are ``print'' statements,
        !           157: generating the output code.
        !           158: .PP
        !           159: The semantic routines that generate the actual 
        !           160: .UC TROFF
        !           161: commands can be changed to accommodate other formatting languages
        !           162: and devices.
        !           163: For example, in less than 24 hours,
        !           164: one of us changed the entire semantic package
        !           165: to drive 
        !           166: .UC NROFF,
        !           167: a variant of
        !           168: .UC TROFF,
        !           169: for typesetting mathematics on teletypewriter devices
        !           170: capable of reverse line motions.
        !           171: Since many potential users do not have access
        !           172: to a typesetter, but still have to type mathematics,
        !           173: this provides a way to get a typed version of the final output
        !           174: which is close enough for debugging purposes,
        !           175: and sometimes even for ultimate use.

This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.