[Haskell-cafe] How Albus Dumbledore would sell Haskell

R Hayes rfhayes at reillyhayes.com
Wed Apr 25 23:06:14 EDT 2007


I don't know how many of the other people on this list are actually  
going to *be* at OSCON.  I will.

I think it is important to think about the kinds of problems the  
audience is trying to solve, as well as the context in which they are  
trying to solve them.  For the most part, the attendees at OSCON work  
on system software: operating systems, operating system services  
(printing, packaging, scheduling, etc), network & distributed  
computing infrastructure, databases, languages, and every imaginable  
variation on the word web.

1) Build a simple database access layer that is immune to SQL  
injection attacks from user input (using the type system to guarantee  
that safety).  Include the FFI portion (FFI is a point of paranoia  
about any "new" language).  Now the, the audience values the type  
system.

2) Show something appallingly simple, yet blazingly fast.   
Demonstrate that it is blazingly fast.  I would implement "wc -l"  
using data.Bytestring.lazy and run in over a huge file.  It's a one  
line program and it will run faster than the built in unix command.   
Explain that its fast because of GHCs rewrite rules.  Explain that  
the rewrite rules are only possible because of purity.   Now the, the  
audience values purity.

3) Build a simple combinator framework that supports multiple  
evaluation schemes (like your derivative framework that supported  
both valuation and settlement processes).  For this audience, it  
might be cool to build a simple package system that both installed  
software and later verified the integrity of the installation.   
Although, if you *wanted* to present "How to write a financial  
contract" I could not object.  Quantitative finance is my business  
and I promise to be in the audience for this.

Also, see if Audrey Tang is going to be present.  If she is, consider  
enlisting her support.  Haskell has *enormous* credibility in the  
Perl 6 community because of PUGS.  I recently had a conversation with  
Jesse Vincent, the Perl 6 program manager.   He said that the  
majority of the Perl 6 community is aware that PUGS (and therefore  
Haskell) are crucial to the success of Perl 6.  Perl has its own  
track at this conference.

reilly hayes




More information about the Haskell-Cafe mailing list