![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to e6 CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [CSRG BSD Unix] / 43BSDReno / share / doc / usd / 26.eqn|
Return to e6 CVS log | Up to [CSRG BSD Unix] / 43BSDReno / share / doc / usd / 26.eqn |
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.