Gee. Isn't predicting the future fun!!
In my eighteen years of programming, I have watched with interest emerging
new technologies and the predictions and expectations surrounding
them. All us tech people, have jumped on various new tech bandwagons as
they go by. We all have competing views on their effect and persistence in
the future.
One problem I noticed is that some of us jump on these bandwagons with the
fervor of religious zealots. We evangelize that the new tech will save the
world, down with the old tech. The only problem with doing this is that we
then fall into this logic-free zone where we aren't free to accurately
evaluate the real pros and cons. We then are married to that tech and thus
lose focus on the larger picture and evangelize just that one thing.
I mean no disrespect to anyone's religion. Earnest belief may have it's
place in religion, politics or philosophy, however, we should give it no
quarter in technology.
I wish to draw some parallels here with a "new tech" of the eighties. In
the mid-eighties some friends of mine started working for a firm using a
4GL (Dataflex) for business systems instead of the 3GLs like COBOL, Pascal
and C. It had "paint the screen" drag and drop, with a powerful language
and a built in DB. Genuinely cool! There were claims that projects in
Dataflex were completed 20 times faster than in COBOL or C!!! WOW! As a
demo I created a simple data entry program (add, change, delete to a table)
in 7 or 8 minutes. The same project in COBOL took me close to 3 hours
using a pre-existing template. This was indeed a 20 to one ratio! And
Dataflax and 4GL was (and is) very cool.
But something was wrong with the claim that projects would be 20 times
faster. I just couldn't put my finger on it. Only after several years of
watching what occurred did it dawn on me what the flaws to this claim were:
1. Projects follow roughly a 40-20-40 rule. 40-50% analysis and
design. 20% coding and 30-40% testing, installation, training, conversion,
etc. Now if coding really is 20 times faster that only affects the
20%. Now a 100 hour project will take 81 hours, not 5 hours!! This is a
~20% improvement. Not 2000%!!!
2. On closer inspection the 20 to 1 ratio only held up on simple
programming. Complex programing with sophisticated conditions and issues
were only somewhat faster coding using Dataflex, not 20 times.
Initial projects they did for clients where the bid was based on the 20 to
1 ratio made some unhappy (and broke) programmers because they were paid on
fixed bid as well. I am not knocking Dataflex or 4GLs. Just the hype some
people fell into.
The outcome 15 years later: 3GLs still dominate in legacy code (# lines of
code worldwide). Much new code is object oriented like Java, C++, Delphi
and others. And the IDEs for these have the drag and drop 4GL features for
fast development. The point is that 4GL was and is great technology, but
was not the end-all solution to all tech problems, the ozone layer and
world hunger.
The parallels to XML? XML like 4GL is a great technology. Turning the
Internet from a world-wide repository of data into a world-wide database is
a very real and exciting prospect. Searching documents with the "smart"
tags of XML will be greatly superior to HTML searches.
But, Please, please, please let us not go the way of the close minded
zealot where we consider using XML format for our EDI needs. I am very hard
pressed to find any real world "value added" to XML despite the
hype. Claims of XML/EDI project speed fall into the same 40-20-40 trap as
above. Thus claims of appealing to the SME (small and medium sized
enterprises) falls flat.
Anthony Beecher wrote: "Is it simply enough to "know better" when the tides
of money and hype are going against you?" No! It is better to be vocal,
even if you are the only little boy in the crowd walking up the big
imposing Emperor and saying "Hey, mister! how come you're naked?" :)
Anthony Beecher wrote: "are any Executives going to take seriously the
information presented in the article?" Unfortunately yes. Executives are
not techies, and will follow trends once they seem to have enough motion
and support behind them. The current motion and support is fueled by
techies excited about XML and it's prospects. This is valid enthusiasm
where XML applies to the web, invalid in my view, where this applies to
business document transactions (EDI).
Gene Hockemeyer wrote: "please note .. it says .. web . second .. it does
not relate this to the total number of B2B transactions from all
sources" Good point. from that perspective it may well turn out to be an
accurate statement.
Cheers,
Steve
------ XML/edi Group Discussion List ------
Homepage = http://www.XMLedi-Group.org
Unsubscribe = send email to: [EMAIL PROTECTED]
Leave the subject and body of the message blank
Questions/requests: [EMAIL PROTECTED]
To receive only one message per day (digest format)
send the following message to [EMAIL PROTECTED],
(leave the subject line blank)
digest xmledi-group your-email-address
To join the XML/edi Group complete the form located at:
http://www.xmledi-group.org/xmledigroup/mail1.htm
