Back from OSCON 2006

» 29 July 2006 » In Books, PHP, Talks »

Just got back from OSCON which was again in Portland this year. The conference was excellent, as always and so were the events and extracurricular activities. The sheer variety of talks at OSCON is exciting and frustrating at the same time: exciting because I attended several talks that I would not get to hear at a more focused conference, and frustrating because of the time conflicts between these talks.

The slides from my own session on PHP 6 and Unicode are online now.

By the way, if you like books just a tiny little bit and happen to be in Portland, do yourself a favor: set aside a full day and visit the Powell’s. It the world’s largest independent used and new bookstore (covering, oh, a couple of city blocks) and has an amazing collection of books (including some very rare ones). You could literally lose your friends and family there and wonder among the stacks for hours whilst salivating giddily over the titles on whatever topic your mind can imagine. And don’t worry, there is always the coffee shop to come back to and get provisions to sustain yourself.

Trackback URL

  1. andrei
    Ryan
    29/07/2006 at 10:59 pm Permalink

    Wow, the transliteration support is AWESOME.

  2. andrei
    Abu Hurayrah
    31/07/2006 at 8:29 am Permalink

    Excellent presentation, Andrei! I’ve got such a load of questions…hope you don’t mind.

    First off, is it outlandish to think that there may be support in the future for Unicode-style built-in PHP functions, classes, & keywords? For example, an extension might define some functions or classes in Japanese, Arabic, Russian, or whatever? I imagine one crazy implementation of this might be a version of PHP entirely localized on one particular language – PHP-Arabic/Hebrew/Russian anyone?

    Secondly, regarding the Unicode error modes, since there are several ways to handle them, is there a “recommended” way or paradigm that the PHP-internals team might propose such that a standard handling method may be developed in the future? By this, I mean should we opt for one particular way that is the most flexible to provide for automated handling? Maybe that’s a little vague, but I hope my intent way conveyed…

    I’m glad to see TextIterator made it into your presentation, and I was pleasantly surprised to see the new Collator object as well – PHP6 really seems to be pushing the OO-envelope (at least for the Unicode extensions…)

    My final comment is regarding transliteration – is this, in any way, customizable? For example, I personally know of a few projects that work with Arabic, and each project adopts their own transliteration technique (e.g., single-character mapping, do long vowels map to multiple characters, use latin diacritical marks, etc.) So, in that vein, is the transliteration procedure fixed, or is it in any way flexible, such as specifying the behavior – or, even better yet, allowing you to provide your own mapping/translation table.

    Also, are any of your talks (this last one in particular) available as free audio or video? It’s be great to see this entire presentation as given.

    Keep up the good work!

  3. andrei
    Andrei
    31/07/2006 at 9:19 am Permalink

    | First off, is it outlandish to think that there
    | may be support in the future for Unicode-style
    | built-in PHP functions, classes, & keywords?

    I sincerely doubt we will ever localize language keywords or standard functions. The new extensions, however, may potentially be able to register Unicode function or class names, but that is not available yet. The Unicode identifier support currently works only with userland stuff.

    As for Unicode error modes, the default is the substitution mode and that would be the preferred one unless you desire to be really strict and just stop processing things on unmapped chars.

    | My final comment is regarding transliteration –
    | is this, in any way, customizable?

    The str_transliterate() function uses built-in transliteration mappings. There will be a more extensive interface to text transforms (OO, most likely) that will allow you to use custom transliteration rules.

Additional comments powered by BackType