Document
maintenance is as expensive, if not more so, than program
maintenance. Isn't is nice to have the same automated
software process to reduce the cost of both?
One of the reasons I stress the need for us to use open source
tools, specifically the editor and compiler, is to be able to
modify them away from file managers to a database manager.
You see you only have two items to deal with, data referents
and data references. You can't reference data without giving
it a name, e.g. "dcl sam fixed binary (31);" in PL/I or "int sam;"
in C. So the software during semantic analysis which seeks
out the uses of such names has no trouble in using it as a
unique identifier for a row in a table. To avoid the one
gotcha in synonyms (same name, different referent) the
software can automatically append an index value to insure
that the combination of the two fields, proper name and
index, are unique.
That takes care of all the referents, whose names are
provided by the programmer (or writer). That leaves only the
references. As all the references are text-based we need
only use some initial prefix of the reference text, say sixteen
bytes, plus an index to generate a unique name for each
reference statement (code) or sentence (text). Again the
software can automatically do this.
Now if you write something which has already been written
before and exists in the source database, the software
through this naming process will determine from the entire
body of the statement or sentence if it already exists. If it
does, it will use the existing source name. Otherwise it will
create a new source entry.
The point is that you no longer has to concern yourself with
reuse. The software will do it for you automatically. It also
means that while you can examine existing source for possible
reuse, a somewhat expensive use of time, you do not have a
need to do so. If it is true for statements and sentences, it is
also true for their assemblies, statement groups and
paragraphs. It holds true for all higher level assemblies.
Now a UNIX programmer may extol the virtues of AWK and
other utilities on processing files. Therein lies the difference.
Files is plural. Database is singular. A singular file is not
global, while a database is. The ability to globally maintain
source regardless of use whether in source code or text or
their combination allows a level of productivity heretofore not
achievable in a file-base system with a file manager without a
builtin query facility.
Thus the fact that editors, compilers, make, linkers, etc.
expect files on input means we are knowingly retaining a less
productive means of processing. Once you integrate your
editor with a database interface such that it automates the
storing, retrieving, maintaining, and reuse of source code and
text at the statement and sentence level, you are closer to
the point where you can implement global changes in the
solution set as fast as they occur in the problem set: dSS/dt
>= dPS/dt.
So we are talking about a 20 byte unique name, a 16 byte
proper name appended with a 4 byte index value, in which we
only have to create ordered assemblies of names to create
documents of source code or source text or any combination
thereof.
So it's not one document, but one document source.
=====================================================
To unsubscribe from this list, send an email message
to "steward@scoug.com". In the body of the message,
put the command "unsubscribe scoug-programming".
For problems, contact the list owner at
"rollin@scoug.com".
=====================================================
>> Next Message >>
Return to [ 23 |
January |
2003 ]
The Southern California OS/2 User Group
P.O. Box 26904
Santa Ana, CA 92799-6904, USA
Copyright 2001 the Southern California OS/2 User Group. ALL RIGHTS
RESERVED.
SCOUG, Warp Expo West, and Warpfest are trademarks of the Southern California OS/2 User Group.
OS/2, Workplace Shell, and IBM are registered trademarks of International
Business Machines Corporation.
All other trademarks remain the property of their respective owners.
