Memoirs of a Roadie

Now that I'm looking to another year of the Birmingham.pm World Tour, with visits to a number of UK LUG and Perl Monger groups, LUGRadio Live (UK not US), the UKUUG Spring Conference in Birmingham, YAPC::NA and YAPC::Europe, as well as possibly a few European Workshops too, I need to start think what I'm going to present. I like the fact I can go to Linux based groups and conferences and talk about a variety of Perl topics, as although I might not be an expert, I know enough to give an introduction in several areas at least. But for more Perl specific technical events, I really need to stick to what I know.

The problem is that I feel I've done enough with CPAN Testing, Phrasebooks and Selenium for the time being, and it does get a bit boring for both me and the audience if I'm repeating myself every year. I may do some update on CPAN Testing, as there are likely to be changes in the coming year, a lot of which is being worked on currently, but what else is there that I could present that would be of interest to somebody?

One talk subject that has crossed my mind has been to do something like 'Labyrinth - A Perl Success Story'. It's been commented a few times that within the Perl community we talk a lot about the possibilities (particularly with frameworks) rather than getting to the finished product. While Labyrinth might not be for everyone, it might possibly be something that works for some, and as a consequence might interest people who have been asking me what it is and why I wrote it. However, although it is related to web and content management it isn't the next Catalyst or the new Jifty. You might be able to draw similarities between them all, but there are also many differences. Labyrinth isn't a framework as such, it's not meant for high-availability websites, and it also doesn't have the large development team knocking out code and fixing bugs that the others have. It's just me. But it might have just enough functionality and usability for someone to pick it up and get a site running how they want it to work, without having to understand the magic internals of frameworks like Catalyst and Jifty. I wouldn't be talking about the internals anyway, as I would prefer to give examples of how I solved problems and interesting asides that led me to learn something new about web design. I'm just not sure enough people would find it that interesting.

Further topics that come from the guts of Labyrinth, and are things that I have been keen to see how other people solve the same problem, are user input validation and content output correction. At the moment Labyrinth handles these within the same codebase, and it works rather well. However, it seems rather the wrong thing to do, to present a talk where the code to do the job isn't on CPAN and is embedded in another system. As a consequence I've been thinking about abstracting the code out of Labyrinth and releasing it separately. It might make for an interesting discussion and may provide people with an reasonable example of how they can use one solution to treat their input and output.

I've also started thinking about doing a short talk along the lines of "My Favourite CPAN Modules". A number of people have done this in the past and at one London.pm meeting several years ago, Leon presented one that got me looking up a few modules I'd not really heard of before. It's probably a talk better aimed at local group technical meetings and maybe a Workshop if appropriate, but I've also been thinking it might be better to actually to structure several talks of this style, but with a theme. So one talk would be "The Web Edition" and feature several modules useful for website development, another "The Test Edition" feature several useful Test modules, and perhaps also "The Mail Edition" with a selection of useful email modules. I've made an attempt at this style of talk before, but got too involved with the mechanics, when really all you need is a quick flavour of what the module can do, with enough references for you to go and find out more yourself.

I still need something more concrete for LUGRadio and the YAPCs, but at least I have some ideas to work with now. If anyone has other suggestions, please let me know.