conversation with zby in #perlcms

(2009-07-10 10:43:05) #perlcms: mode (+nt ) by magnet.shadowcat.co.uk
(2009-07-10 10:49:50) zby [~zby@19-tor-3.acn.waw.pl] entered the room.
(2009-07-10 10:50:05) robinsmidsrod: hopefully Theory at some point
(2009-07-10 10:50:07) zby: heh - I see
(2009-07-10 10:50:09) robinsmidsrod: (David)
(2009-07-10 10:50:14) zby: OK
(2009-07-10 10:50:31) zby: anyway first the problems I see
(2009-07-10 10:50:36) robinsmidsrod: in fact, I'm thinking that maybe we don't need to actually choose a framework
(2009-07-10 10:50:39) zby: I think you need to choose a platform
(2009-07-10 10:50:45) zby: Mojo is interesting
(2009-07-10 10:51:00) robinsmidsrod: maybe we can just make a dispatching plugin for any of the large ones
(2009-07-10 10:51:02) zby: but you'd need to find more people that use it
(2009-07-10 10:51:11) zby: I have no experience with it
(2009-07-10 10:51:21) robinsmidsrod: me neither - I have just skimmed the docs
(2009-07-10 10:51:26) zby: hmm 
(2009-07-10 10:51:36) robinsmidsrod: in fact, I'm not that experienced with catalyst either
(2009-07-10 10:51:39) zby: I am now experimenting with having the dispatcher in the controller
(2009-07-10 10:51:40) robinsmidsrod: I've just made one app with it
(2009-07-10 10:51:48) robinsmidsrod: actually, two apps
(2009-07-10 10:51:55) zby: so in theory you could reuse it in other framework
(2009-07-10 10:51:58) robinsmidsrod: one is here: photo-competitions.net
(2009-07-10 10:53:04) robinsmidsrod: because the only thing that is really framework-relevant is how to map the request/response objects and how to dispatch paths
(2009-07-10 10:53:21) robinsmidsrod: maybe it would be bound a bit more when we approach auth/authz
(2009-07-10 10:53:28) zby: Caught exception in ->ACL error rethrower "Access to competition/browse/all/index denied by rule CODE(0x470ed60) (defined at /home/vhosts/photo-competitions.com/script/../lib/PhotoComp.pm line 67)."
(2009-07-10 10:53:53) robinsmidsrod: guess I need to fix that 
(2009-07-10 10:54:01) zby: yes - the dispatching is the most framework dependant
(2009-07-10 10:54:08) zby: but there are other things
(2009-07-10 10:54:24) zby: like the upload field
(2009-07-10 10:54:29) zby: in the request object
(2009-07-10 10:54:39) zby: small things - but they crop up everywhere
(2009-07-10 10:56:12) zby: then there is auth
(2009-07-10 10:56:46) zby: I think you need to choose the framework
(2009-07-10 10:56:50) zby: to start with
(2009-07-10 10:57:07) zby: you need to write a prototype to show people
(2009-07-10 10:57:10) zby: and attract developers
(2009-07-10 10:57:47) robinsmidsrod: zby: yeah, but first I need to think about the data/dispatch model I want to use
(2009-07-10 10:57:57) robinsmidsrod: now that site works again
(2009-07-10 10:58:15) robinsmidsrod: still problems...
(2009-07-10 10:59:15) robinsmidsrod: it's just an uncaught exception because I didn't remove an url that was supposed to be hidden for guest users - it was a school project - so it's not exactly well-polished
(2009-07-10 11:00:41) zby: OK 
(2009-07-10 11:00:57) zby: I've skimmed your github readme
(2009-07-10 11:00:58) robinsmidsrod: well, it's my biggest catalyst project though
(2009-07-10 11:01:22) robinsmidsrod: what are your immediate thoughts?
(2009-07-10 11:01:29) zby: what is the difference between path and block?
(2009-07-10 11:01:40) robinsmidsrod: a path is basically an URI
(2009-07-10 11:01:52) robinsmidsrod: think of it like a key to a list of blocks
(2009-07-10 11:02:37) zby: OK - but blocks are also list of blocks?
(2009-07-10 11:02:38) robinsmidsrod: a block is a small part of the entire page that will actually render that path
(2009-07-10 11:03:04) robinsmidsrod: zby: that's the part I'm not sure about - if we should allow blocks to wrap other blocks
(2009-07-10 11:03:10) robinsmidsrod: or if it will just complicate stuff
(2009-07-10 11:03:34) zby: I would propose that it is recursive
(2009-07-10 11:03:53) robinsmidsrod: zby: yeah, that gives us much more flexibility
(2009-07-10 11:03:54) zby: but I would say that a path identifies a block
(2009-07-10 11:04:44) robinsmidsrod: hmm...interesting
(2009-07-10 11:05:02) robinsmidsrod: and then you just assume that a block, if it is static just wraps to the entire page
(2009-07-10 11:05:13) robinsmidsrod: if it is dynamic it probably includes children
(2009-07-10 11:05:42) zby: and you could also render sub-blocks
(2009-07-10 11:05:48) zby: because they would have paths
(2009-07-10 11:05:58) zby: this could be usefull for debugging
(2009-07-10 11:06:26) robinsmidsrod: rephrase that thing about sub-blocks
(2009-07-10 11:06:32) robinsmidsrod: I didn't quite get it
(2009-07-10 11:07:10) zby: you let a block contain other blocks
(2009-07-10 11:07:29) zby: i.e. sub-blocks
(2009-07-10 11:07:36) robinsmidsrod: now I remember why I was thinking that a path is a list of blocks - it's to be able to work nice towards ESI
(2009-07-10 11:07:40) robinsmidsrod: (Edge Side Includes)
(2009-07-10 11:07:56) robinsmidsrod: to allow static and dynamic blocks in the same page
(2009-07-10 11:08:18) zby: yeah - so I would say that a path leads to a block
(2009-07-10 11:08:27) zby: and that block can contain the list of sub-blocks
(2009-07-10 11:08:39) zby: this would be nicely recursive
(2009-07-10 11:09:00) robinsmidsrod: okay, how do we map this so that ESI can get to it?
(2009-07-10 11:09:15) zby: I have no experience with ESI
(2009-07-10 11:09:24) zby: give me some link - so that I'll read about it
(2009-07-10 11:09:45) zby: anyway - now I need a daughter emmergency break
(2009-07-10 11:10:02) robinsmidsrod: maybe this one? http://varnish.projects.linpro.no/wiki/ESIfeatures
(2009-07-10 11:11:00) robinsmidsrod: I don't know a lot about ESI either, but I know it's used a lot in high-performance setups
(2009-07-10 11:11:12) robinsmidsrod: and I'd like to accomodate for it in the design
(2009-07-10 11:11:54) robinsmidsrod: even better link: http://en.wikipedia.org/wiki/Edge_Side_Includes
(2009-07-10 11:14:25) zby: OK - thanks - I'll read about it
(2009-07-10 11:14:45) zby: but to give you some perspective on what I work now on
(2009-07-10 11:14:58) zby: I have in mind a Catalyst::Elements project
(2009-07-10 11:15:11) zby: for assembling basic features to a Cat application
(2009-07-10 11:15:17) zby: like authentication
(2009-07-10 11:15:20) zby: comments
(2009-07-10 11:15:23) zby: etc
(2009-07-10 11:15:31) zby: CRUD
(2009-07-10 11:15:53) zby: I think this is now my priority
(2009-07-10 11:16:22) zby: but I can take part in the discussions about the cms
(2009-07-10 11:16:44) zby: and if you have any particular needs - then I can try to write some Element for it
(2009-07-10 11:16:53) zby: what do you think about that?
(2009-07-10 11:17:03) robinsmidsrod: let me think before I answer
(2009-07-10 11:17:12) zby: yeah - sure
(2009-07-10 11:18:03) robinsmidsrod: your concept seems smart too, the concept being to assemble catalyst components like legos into forming a complete cat-app
(2009-07-10 11:18:32) zby: yes - this is actually quite obvious one
(2009-07-10 11:18:53) zby: what I add to it is that I am trying to make it simple and replaceable
(2009-07-10 11:18:54) robinsmidsrod: it's not so far from what I have in mind - 
(2009-07-10 11:18:58) zby: scaffolding
(2009-07-10 11:19:14) zby: yeah - I hope so :)
(2009-07-10 11:19:45) robinsmidsrod: but I guess I'm more focused on getting the foundation right so that caching and expiry and plugins work with very little effort
(2009-07-10 11:20:18) robinsmidsrod: actually, I somewhat see this as a refinement to what I do at work, but more generic - see portfolio.no
(2009-07-10 11:20:52) robinsmidsrod: that is a project focused on the needs of learning
(2009-07-10 11:21:02) robinsmidsrod: web-based learning
(2009-07-10 11:21:27) zby: interesting
(2009-07-10 11:21:28) robinsmidsrod: so I'm trying to take what knowledge I have from that project and generalize it into 
(2009-07-10 11:21:37) robinsmidsrod: I've been working on that thing since 2005
(2009-07-10 11:21:59) robinsmidsrod: at that time catalyst was just in its infancy (at least from my perspective) so it's not based on catalyst
(2009-07-10 11:22:30) zby: yeah catalyst started in 2005
(2009-07-10 11:23:22) robinsmidsrod: so you can probably understand where the ideas I've written down come from
(2009-07-10 11:23:43) robinsmidsrod: it's sort of a refinement of the stuff I've been doing at work for the last 4 years
(2009-07-10 11:23:52) zby: great!
(2009-07-10 11:24:20) zby: OK - so let me read about the ESI
(2009-07-10 11:24:46) zby: and think a bit how could you fit my Elements idea into your project
(2009-07-10 11:26:21) robinsmidsrod: the very simple concept is that building a website should be like constructing it from lego bricks
(2009-07-10 11:26:54) robinsmidsrod: that's how I thought of the term "block"
(2009-07-10 11:30:07) zby: :)
(2009-07-10 11:32:45) robinsmidsrod: we also need to think about versioning and deprecation of features, and APIs, to make sure that upgrading and such is also clean
(2009-07-10 11:34:00) robinsmidsrod: I was thinking that the blocks themselves could be version, something like what git does with blobs
(2009-07-10 11:34:11) robinsmidsrod: s/version/versioned/
(2009-07-10 11:34:26) robinsmidsrod: this could also enable wiki-like sites that allow rollback of content and history
(2009-07-10 11:35:02) robinsmidsrod: and then maybe a system to purge unreferenced blocks or something
(2009-07-10 11:35:48) robinsmidsrod: as you can see, I have lots of thoughts on the subject - but I'm trying to take a minimalist approach to what the core system should handle
(2009-07-10 11:36:37) robinsmidsrod: I don't know if it's approriate, but think of it somewhat like git for web-development with dynamic content (generated from perl code) thrown into the mix
(2009-07-10 11:37:02) robinsmidsrod: maybe even some code signing for authenticated plugins/blocks or whatever
(2009-07-10 11:40:26) robinsmidsrod: as a photographer and programmer I think of useful applications as e.g. a blog where I talk about programming - my portfolio of my pictures, my photo-plog (which is mostly pictures) - all of this should be possible to host on one installation of the system
(2009-07-10 11:40:46) robinsmidsrod: but I'm also thinking that it could be useful as a ISP hosting solution aswell
(2009-07-10 11:41:38) robinsmidsrod: so a strong API for manipulating the data structures via REST/SOAP or something sounds about right
(2009-07-10 11:42:05) robinsmidsrod: this way you could make tools like the flickr uploader that would allow you to keep your portfolio site updated offline
(2009-07-10 11:42:35) robinsmidsrod: or you could keep a staging system at one location and just sync it to production when you're ready
(2009-07-10 11:42:45) ***robinsmidsrod is just throwing out ideas
(2009-07-10 11:44:47) zby: :)
(2009-07-10 11:44:54) zby: thats a long list
(2009-07-10 11:52:20) robinsmidsrod: yeah - it is
(2009-07-10 11:52:28) ***robinsmidsrod is eating - be back soon
(2009-07-10 13:14:26) robinsmidsrod: okay, I'll be off to do some personal errands, but if you have anything to share about the project please send a pull request for the docs, give me a patch or just shout of some ideas here - I'll eventually be back and will read them
(2009-07-10 13:14:30) robinsmidsrod: c u
(2009-07-10 13:14:53) robinsmidsrod: if Theory comes online, try to invite him here...
(2009-07-10 13:15:13) robinsmidsrod!~robin@197.84-49-220.nextgentel.com: robinsmidsrod has changed the topic to: http://github.com/robinsmidsrod/unnamed-perl-cms-project
(2009-07-10 13:15:45) robinsmidsrod!~robin@197.84-49-220.nextgentel.com: robinsmidsrod has changed the topic to: http://github.com/robinsmidsrod/unnamed-perl-cms-project - have a look at the readme and give me your input
(2009-07-10 18:39:26) Theory [~Theory@c-76-105-135-51.hsd1.or.comcast.net] entered the room.
(2009-07-10 18:40:12) robinsmidsrod: I had a discussion with zby earlier today
(2009-07-10 18:41:08) robinsmidsrod: Theory: I'm not sure how this aligns with your concept - and btw - why are you interested in this when you obviously are deeply involved in bricolage? is it not doing what you want?
(2009-07-10 18:43:28) Theory: Bricolage is a back-end workflow and publishing system.
(2009-07-10 18:43:32) Theory: It doesn't actually serve sites.
(2009-07-10 18:43:46) Theory: Working on a project now where $client develops content in Bricolage and publishes it to a Drupal server.
(2009-07-10 18:43:52) Theory: Drupal has filled me with hate.
(2009-07-10 18:43:56) Theory: QED
(2009-07-10 18:44:48) robinsmidsrod: okay - I have some logs from my conversation with zby, so that you have a bit more meat on the bone in addition to the README
(2009-07-10 18:44:58) robinsmidsrod: btw, please clone that repo and add your bits