Profilin' and Serializin'

samtregar on 2002-10-17T02:59:47

My work on DBI::ProfileDumper has reached its first public milestone; the patch is in. I'd be foolish to claim that my work is done, but I will proudly state that it works for me.

This frees me up to return to a problem I'd put on the back burner - serialization. To put it plainly, I'm sick of XML and I'm not going to take it anymore. Using XML as a serialization format for Perl data-structures stinks. It requires mountains of custom code to support, produces outrageously bloated output and makes dealing with 8-bit ASCII data needlessly difficult. XML simply wasn't designed as a data serialization language and it shows.

Enter YAML, a programming language specifically designed to serialize data structures in a human readable and editable form. The format is compact, beautiful to look at and includes native support for the stuff that matters (scalars, hashes, arrays and objects).

Unfortunately, it appears that Perl's YAML implementation is in need of some work. It's not compliant with the current spec and it has a few fairly significant deficiencies. At this point I'm reading the spec, surfing the YamlWiki, chatting up the YAML crew and waiting for access to the YAML dev tools. I'm hoping to massage YAML.pm into shape in time for Christmas (or the start of my next project, whichever comes first).

-sam


YAML

Theory on 2002-10-17T15:13:24

I think that Brian is working on getting YAML.pm up to the current spec, now. Not positive, though.

--David

Re:YAML

samtregar on 2002-10-17T15:17:52

Yeah, we've been in contact. I started to include that fact in my entry but I thought it was too name-droppy for some reason. I suppose that's pretty funny; I'm pompous enough to keep a public journal but too shy to name my friends in it!

-sam