Finishing with the open-versus source debate.

charles at charles at
Sun Jun 23 00:31:48 EDT 2002


you bring about a very good point. 

What have WE done for or with open source. What is the measure of our

If we can ever get the project to the next step, I hope to provide a
sound O-O footing for OpenBiblio.

Some history:

I also hope to provide some enhancements to phpwiki with some dialog
and transaction processing and state machine/transition engine

When I started in Smalltalk in the late eighties until the late
nineties when Digitalk found a way hide the source code ALL my code
was open source. Smalltalk products didn't ship as anything BUT
because they have to be integrated into an image file which the VM
then loads and runs with.

This led to some tighter and neater code. You KNEW everybody would be
seeing it.

Code I wrote that was used in a course I tried to give on Compuserve.
That was simply too early in my own evolution and in Smalltalk's
acceptance (more the failure thereof,) to get widely used or

Code I wrote for converting TIFF and GIF images to/from images that
were internal to Smalltalk.

Code I wrote for convolving images (edge detection, shade smoothing,
contrast enhancing and other image manipulation code,) was all open

Code I developped for the Government of Canada (expert systems,
expert applications, some in Smalltalk and some using "standard"
expert system shells a.k.a. inference engines.)

Code I wrote for and wrote about in Smalltalk report, Byte Magazine,
Computer Language Magazine, AI Expert.

Code I wrote for various other projects was all open source and the
client got the source along with the product.

Actually, it wasn't until I hit the United States in '95 that people
began locking up my code and tying my tongue with confidentiality
agreements. That's also when I started seeing code that was
inexcusably lousy, amateurish and fragile.

The project's I had worked on before had had some pretty poor code
but nothing so universaly lousy as what was being developped and
"shirk"-wrapped for general consumption. I'm ashamed of some of the
comments I have written and read in commercial code or that the code
itself was so sloppy.

When I was working in Montreal in 1984, I had to maintain an
accounting package written by McCormack and Dodge. It was in COBOL
and we had the source code. We'd find and fix bugs and submit them
for inclusion into the package and our efforts were apppreciated and

In fact, NOBODY who's going to lay out significant money ($100k+) for
software is NOT going to get the source code. Its NOT happening. You
may have to sign non-competition agreements but you're getting the
source code or they're NOT making the sale.

Likewise, you're NOT getting code on a mainframe unless you have the
source code and it gets subjected to "code efficiency" peer-review,
process "weight" aanlysis for performance metrics and capacity
planning, data base access method (DBAM) review and a complete
security audit. Your code is not getting installed unless they can
compile the source with their own compilers. Its NOT happening.

We in the PC arena seem to have problems and issues that were solved
decades ago on mainframes. Its only the PC arena which refuses to
listen and arrogantly assumes that nothing will go wrong with the
code when all evidence gathered from all experience in everything
else is that "Shit Happens!"

The help desk advice of "re-install and retry" would get you a ticket
on a bus in a mainframe shop.



Some MORE history:

Closed-source did not exist at all as a concept until a whiney Bill
Gates published something in Byte magazine in the mid-to-late
seventies. Aactually, he bought an ad, ranting that people were
ripping off his BASIC interpreter (that he himself had ripped off
without paying royalties of acknowledging the originators) and it was
costing him money in lost sales. That's who closed source has been
good to. He's worth hundred's of billions of dollars and you're worth

Closed source doesn't exist except on PCs. ("Big Iron" get the source
and compiles it with its own trusted compilers in a series of
development, test, QA, Integration and, finally, production

Its called trying to sell me a pig-in-a-poke and it ain't happening.
It shouldn't happen to anybody.

More information about the talk mailing list