Re: [office-comment] OpenFormula - scope, other info resources?

["David A. Wheeler" <dwheeler@dwheeler.com>]

I said:
>>* http://sc.openoffice.org/ and in particular
>>  "OpenOffice.org's Documentation of the Microsoft Excel File Format"
>>  is an absolute delight for real, meaty technical details.
>>  It's also scary; it's unbelievable that so many people
>>  use spreadsheets for exchanging information,
>>  even though there's no official spec and the Excel spreadsheet
>>  format (as documented here) is INCREDIBLY fragile....

John Cowan replied:
> Well, a quick browse through that was most enlightening.  Evidently
> this format was designed for slow (floppy-based?) computers, where
> there was a premium on loading and saving as quickly as possible,
> so it's essentially a core dump.

Yes, I think you're exactly right, it's a core dump.

But that doesn't mean it's a good trade-off today.
This structure is extremely fragile and hard to expand, and
people are willing to do things differently to increase
reliability and flexibility.

I haven't done the benchmarks, but I suspect that there
may actually not even be much of a trade-off... they
probably decided to do what was easy, not what was best.
Even when 4.77MHz chips were the norm, the floppy drives
were MUCH slower than the CPU... I suspect uncompressing
and parsing (and its inverse), which are CPU-intensive, were
so fast that the CPU would still be faster than the floppy.
And at least for the word processing format, zipped
XML files are actually much SMALLER than the binary
"efficient" format.  It's once again the "do one thing,
do it well" notion -- a program specially designed to
do compression can often mop up the hand-created
sorta-compressed formats.

Today, of course, there isn't even a trade-off.
Reliability & flexibility are far more important.

--- David A. Wheeler