introspection (and Perl 6)

[muppet]

On Jan 24, 2008, at 6:33 PM, jesse wrote:

> Spreadsheets do an amazing job of letting casual end users get useful
> programming done. I keep wondering if building tools around a  
> databasey
> paradigm rather than a spreadsheety one is the right thing.

You can only go so far in the spreadsheety paradigm.  They always say,  
"this is just a one-off", but they don't mean it.  Eventually, someone  
will want to use your spreadsheet prototype for process automation.

Ages ago at $dayjob, i had a cow-orker who spent several months  
writing Fourier transforms and other crazy analysis crap in Excel,  
pushing pivot tables to the limit, with super fancy charts, complex  
non-tabular input file parsing, simply an amazing amount of stuff.   
But the ability to repeat this analysis on other copies of Excel on  
other users' desks, with other datasets, was estimated at "a fair  
amount of work".  Nobody else could get it to work at all.

I and a different coworker reimplemented this whole thing with perl,  
mysql, php, and gnuplot in about three hours.  It has been working  
with truly minimal maintenance for about seven years, now, with tens  
of thousands of data records.  Please note that we were nowhere near  
what you'd call software experts at the time.

In short, the model we discovered was "prototype it in a spreadsheet,  
then port it to Real Tools for process automation."


--
If the monkey could type one keystroke every nanosecond, the expected  
waiting time until the monkey types out Hamlet is so long that the  
estimated age of the universe is insignificant by comparison ... this  
is not a practical method for writing plays.
   -- Gian-Carlo Rota