[13:00] <mitchell> OK, it's right about 1pm, so we'll get started
[13:01] <mitchell> for those folks not from OSAF, if you would like to say hello and let us know if you have any particular interest, now would be a good itme
[13:01] <mitchell> otherwise, we'll just get started in general terms
[13:01] jonathan (~jprusky@12-236-204-253.client.attbi.com) joined #chandler.
[13:01] <mitchell> I'm assuming eveyone knows of a few key documents, but I'll post the references just in case
[13:02] <mitchell> An overview of design values:  http://www.osafoundation.org/architecture.htm
[13:02] <mitchell> Block Architecture diagram:  http://www.osafoundation.org/architecture.htm#block_diagram
[13:03] <mitchell> and textual description of the Architectureal overview:  http://www.osafoundation.org/architecture.htm#architectureoverview
[13:03] <Empty> /me /mode -i
[13:03] <mitchell> OK, so it looks like no one wants to say hello yet
[13:03] <mitchell> except empty, who's clearly struggling
[13:04] <jonathan> hi mitchell
[13:04] <mitchell> So John, maybe you could tell us what you're working on now
[13:04] <mitchell> jonathan:  not sure if you say my earlier comment, suggedting people identify particular architectural interests
[13:04] <john> I'm working on a bunch of things, perhaps the most important is Mitch is "super widget"
[13:05] <mitchell> oh yes, people asked about that last week, and I couldn't say much
[13:05] <mitchell> can you give us the big picture "vision" first
[13:05] Empty (~MichaelT@63.90.156.153) joined #chandler.
[13:05] <mitchell> and then some more detail about implementation thoughts?
[13:05] <john> Mitch owns the vision piece, which is evolving
[13:05] af (~andrewfr@216.13.46.139) joined #chandler.
[13:05] #chandler: mode change '+o Empty' by jurgen!~jbotz@63.90.156.153
[13:06] #chandler: mode change '+o chao' by jurgen!~jbotz@63.90.156.153
[13:06] <john> I'm currently in brainstorming node about different ways to implement some of his ideas
[13:06] <mitchell> can you give us an idea of what the super-widget is supposed to do?
[13:06] Ilan (Gorvok@user-152-16-70-183.adsl.duke.edu) joined #chandler.
[13:06] <john> at its fundamental level, we are beginning to think of views being more like documents
[13:07] <john> where a document is organized in a hierarchical nature
[13:07] <mitchell> and Andy, feel free to join in, if you've been part of this
[13:08] <john> each node is similar to what we've been thinking of as a view until now
[13:08] <john> that is, a list of items (possibly generated by a query) displayed in a particular viewer
[13:09] <mitchell> I think of the super-widget concpetually  as  it as a three-dimensional outliner, allowing items to be organized and manipulated  in many views
[13:09] <Andy> we will allow users to customize the built-in documents (formerly views), and construct new ones
[13:09] <john> these "documents" might be items that could be shared like other items
[13:10] morgen (~morgen@66.159.224.4) joined #chandler.
[13:10] <john> there are still lots of unsolved open issues about the details, but some of the central themes are starting to gel
[13:10] <Andy> sort of like Hypercard, or Interface Builder
[13:11] <mitchell> Andy:  can you say more about "like hypercard"
[13:11] <jonathan> why the switch to "documents" rather than "views"?
[13:11] <Andy> you can instantiate new widgets to view items, and add scripts to specify behavior
[13:12] <Andy> documents seems to feel more tangible than views
[13:13] chao (~chao@63.90.156.153) left irc: Remote host closed the connection
[13:13] <Andy> although the terminology may continue to shift for a while
[13:13] <jonathan> so will there still be "views" elsewhere, and if so, will that take on a different definition
[13:13] #chandler: mode change '+o morgen' by jurgen!~jbotz@63.90.156.153
[13:13] chao (~chao@63.90.156.153) joined #chandler.
[13:13] <luther> john: These documents you refer to are content, presentation, or a bit of both?
[13:13] <john> in the past a view wasn't an item, now a document is, and it's seems more natural to think of authoring the document and sharing it instead of authoring a view
[13:13] <Andy> no, documents will replace what we were calling views
[13:13] <john> documents referred to items
[13:13] Topic changed on #chandler by jurgen!~jbotz@63.90.156.153: Chandler Architecture with John Anderson and Andy Hertzfeld: chat in session
[13:14] #chandler: mode change '+o chao' by jurgen!~jbotz@63.90.156.153
[13:14] <john> documents contain structure and format imposed on a hierarchical tree of lists of items
[13:15] <john> but documents don't contain items, just pointers to items, and queries making up lists of the item pointers
[13:15] <jonathan> so there will be a "Weekly document" which might show you all the appointments you have for the week?
[13:15] <Andy> a document is a bunch of nodes arranged in a hierarchy
[13:16] <john> yes, you could set up such a document if you chose to
[13:16] <Andy> a node is a query (or perhaps an explicit list of items) bound to a widget to display the items
[13:16] <Andy> parcels will come with a bunch of documents
[13:16] <Andy> but you'll be able to customize them, or add new ones.
[13:16] <john> and share them
[13:17] <jonathan> so for now, the term "view" is retired?
[13:17] <Andy> that's our current thinking
[13:17] <luther> So, are we talking something akin to DOM, except with the domain of its documents being the items in the repository (and perhaps elsewhere)?
[13:17] <john> these ideas are preliminary, so nothing is yet etched in stone, but that's where it looks like things are headed
[13:18] <john> yes it's similar to DOM
[13:18] <Andy> yes, it's very similar to a DOM, especially since you'll have scripts that operate
[13:18] <Andy> on the nodes
[13:18] <mdubinko> sounds great, when can we see the code :-)
[13:19] <jonathan> sorry, what is "DOM"?
[13:19] <luther> DOM = Document Object Model
[13:19] <Andy> DOM = document object model
[13:19] <luther> GMTA
[13:19] <jonathan> sorry
[13:19] <Andy> popularized by web browsers
[13:19] <sprout> how DOM like are you thinking here?
[13:19] <pieter> Won't the actual content of a 'document' therefore be dynamic? In that case, how does one communicate between the author(s) and the folks that are 'sharing' this document the particular version since the content you might be referring to might be modified or deleted? 
[13:20] <john> http://www.w3.org/DOM/
[13:20] <Andy> the content could be explicit, or it could be a query
[13:20] <Andy> the queries could be applied to the local repository once it's shared
[13:20] <pieter> with some UI to differentiate static vs. dynamic content?
[13:21] <Andy> yes
[13:21] <john> and of course there's the issue of multiple users modifying the same content, which brings up the usual collision detection and arbitration
[13:21] <sprout> since everything in chandler has a URI, would you also be looking at XPath/XPointer to specify stuff "inside" documents?
[13:22] <Andy> yes, I think xPath/XPointer will be useful in scripting
[13:22] <john> I'm not sure which you mean by "XPath/XPointer"
[13:22] <Andy> they're xml constructs for manipulating xml documents
[13:23] <sprout> XPath and XPointer are extensions for querying XML/HTML documents.
[13:23] <john> thanks
[13:23] <af> and then there is Xquery
[13:23] <Andy> yes, we might use XQuery for our queries, but nothing's been decided yet
[13:24] <sprout> http://www.w3.org/TR/xpath and http://www.w3.org/XML/Linking
[13:24] <sprout> A sizable portion of XQuery is based on XPath
[13:24] <john> how we do queries is a big open issue
[13:24] <af> yes any valid XPATH expression is a valid XQuery expression.
[13:25] <mitchell> Andy, what are you working on these days?
[13:26] <Andy> I'm mainly focusing on the Agent Framework now
[13:26] <Andy> we hope to have some working agents in the .2 release
[13:26] <mitchell> what agents might we see?
[13:26] <Andy> so right now I'm trying to define APIs for the event manager, whose name just changed
[13:27] <Ilan> Please, no talking office supplies. '
[13:27] <Andy> to the notification manager, to avoid confusion with wxWindows events
[13:27] <Andy> the first agent will be an "approval" agent for managing presence subscriptions and other
[13:27] <Andy> situations that require the user's approval - you'll be able to delegate that to an agent.
[13:28] <mitchell> like calendar requests?
[13:28] <Andy> but before we implement specific agents, we need to have some framework machinery in place
[13:28] <Andy> calendar requests will probably be handled by a calendar-specific agent
[13:29] <Andy> so first we're working on notifications/events, and then actions and trigger objects, and then
[13:29] <af> Andy, what would have to be added/changed to Chandler to accommodiate the Agent Framework?
[13:29] <af> i.e (notification manager, concurrency model, etc)
[13:30] <Andy> yes, notifications and a concurrency model are necessary
[13:30] <mitchell> Is this the correct URL for the thinking on the framework machinery? http://wiki.osafoundation.org/bin/view/Jungle/EventModel
[13:30] <Andy> plus some specific UI; I have a first cut on the drawing board
[13:31] <Andy> yes, that's the current thinking for events (to be renamed notifications)
[13:31] <Andy> I'm going away on vacation for a week tomorrow, but hope to start implementing
[13:31] <Andy> in June when I get back.  We should have something running by the end of June.
[13:32] <Andy> af has been helping a lot with defining the APIs for events, actions and triggers
[13:32] jonathan (~jprusky@12-236-204-253.client.attbi.com) left irc: Ping timeout: 180 seconds
[13:32] <mitchell> af:  I had no idea that was you!
[13:33] <af> ?
[13:34] <mitchell> for the non-OSAF folks, now's your chance to ask Andy and John anything that's on your mind about Chanlder architecure
[13:34] <mitchell> i guess, the same is tre for the OSF folks as well :-)
[13:34] <af> meanwhile (and I am a little behind), I am trying to workout what the behavior of the event manager is.
[13:34] <af> rather notification manager.
[13:35] <Andy> so between the new document paradigm, the outline/table widget and the agent framework, the .2 release is shaping up to be pretty exciting!
[13:35] <john> not to mention the database and items
[13:35] <mdubinko> is someone wanted to help out for the .2 release, what would be the best way to proceed?
[13:36] <Andy> it depends on what kind of expertise you have
[13:36] <Andy> we could use all kinds of help at various levels
[13:36] <john> give us feedback on what's hard or inconvenience to do when writing a viewer parcel
[13:37] <luther> john: Lack of good wxPython docs for writing the GUI aspects...
[13:37] <john> that's one of the items I'm currently working on
[13:37] <Andy> someone is writing an O'Reilly book on wxPython
[13:37] <luther> bless you.
[13:37] <luther> Andy: is there a beta on the web somewhere?
[13:37] <john> You'll have to ask Robin
[13:38] <Andy> I think so, Robin Dunn would know
[13:38] <mitchell> Andy, John, can you say much about our repository direction?
[13:38] <Empty> last i heard from robin the wxpython book was pretty early in it's process.
[13:39] <john> yes, the repository has been one of our biggest challenges.
[13:39] <john> We just hired Andi, to take on the task.
[13:39] <robin> yes, and the coauthor (who was doing most of the work) just got a full time gig and so the schedule for the book is up in the air at the moment.
[13:40] <robin> We are however working on getting Python-specific reference docs and such done.
[13:40] <john> the first goa lfor the repository is to make a pass over Katie's cod for creating item schemas and items
[13:41] jonathan (~jprusky@12-236-204-253.client.attbi.com) joined #chandler.
[13:42] <jonathan> is this chat still in session?
[13:42] <mitchell> jonathan:  yes
[13:42] <mitchell> at a somewhat leisurely pace
[13:42] <john> once we have agreement on item schemas, we will work on some new APIs for accessing items, besides the usual transparent persistence
[13:43] <mitchell> john was just describing our repository direction
[13:43] <jonathan> sorry I got disconnected
[13:43] <sprout> is the current plan to stay with the ZODB back end?
[13:43] morgen (~morgen@66.159.224.4) left irc: Ping timeout: 180 seconds
[13:43] <john> the goal is to settle on how applications access items, which will allow us to modify the back and without affecting applications
[13:44] <john> the decision for whether or not to use ZODB is up two Andi
[13:44] <Ilan> Does the current Chandler architecture support dynamically adding parcels (i.e. adding them while Chandler is running)? 
[13:44] <sprout> so it sounds like if there's going to be a query language, that this is part of what needs to happen with the repository soon
[13:45] <Andy> yes, parcels can be dynamically added (although that's not working in .1)
[13:45] <Ilan> So it works in the CVS version?
[13:45] <Andy> yes, we need to get at least a strawman query language going ASAP
[13:46] <john> but in the end, will end up with an object cache, backend database, and a protocol between them
[13:46] <Andy> no, not in cvs yet, either, but that's our intention
[13:47] <af> concerning the notification manager, do you have a particular model in mind? 
[13:47] <Andy> not sure what you mean by model
[13:47] <af> existing implementation or theoretical model.
[13:48] <Andy> no existing implementation, but we're basing our first implementation on wxPython's pubsub.py
[13:48] <Andy> we know that won't be sufficient in the long run, but we'll use it to get started
[13:49] <Andy> you can declare notifications in a hierarchical namespace, discover them, and then subscribe to them and post them
[13:49] <john> do you think the API will change in the future extensions?
[13:49] <mdubinko> There's another XML spec in that space: XML Events
[13:49] <Andy> I'm hoping to keep the API the same, but change the underlying implementation as it evolves
[13:50] morgen (~morgen@66.159.224.4) joined #chandler.
[13:50] <Andy> Do you have a link to xml events?  I never heard of them
[13:50] <mdubinko> http://www.w3.org/TR/xml-events/
[13:50] <Andy> thanks
[13:50] <mdubinko> It's based on DOM Level 2 Events
[13:51] <af> the events in the notification manager are more general
[13:52] <john> in what sort of ways?
[13:52] <Andy> I think they have a higher level of semantics
[13:52] <mitchell> We're coming up on 2pm, so anyone who wants to ask a question should do so soon
[13:52] <Andy> the dom events are things like mouse-clicks
[13:52] <mdubinko> some of the DOM events are really general (like "load", or "activate") and some are rather low level
[13:53] <af> The XML events look like they pertain to XML documents, I think of events pertaining from everything from a mouse click to a Concert Annoucement about a show.
[13:53] <Andy> we're talking about events like "message arrived"
[13:53] <af> events are about occurances of conditions or activities.
[13:53] <af> I use event in the workflow/reactive system sense of the word.
[13:54] <Empty> interesting, shouold the "super widget" expose itself as a DOM-compliant document?
[13:54] <john> I don't know yet
[13:54] <Andy> we need to decide if we're going to embrace the DOM as the scripting API
[13:55] <Andy> it sounds like it might be a good idea
[13:55] <sprout> I'd rather see the super widget accessible via XPath
[13:55] <john> do you have any idea who will decide what the events are?
[13:55] <mdubinko> I've been doing some DOM scripting lately, and it turns out to be nicely compatible with XPath
[13:56] <Andy> the system will define some events, and others will be defined by parcels or agents
[13:56] <sprout> There is a really good article at MSDN (of all places) on using XPath to navigate not just XML documents, but object hierarchies.  
[13:56] <jonathan> andy: do have any design/development timeframe defined for the super-widget.. even a prototype?
[13:56] <Andy> john should answer that
[13:57] <sprout> http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnexxml/html/xml03172003.asp  worth a look
[13:57] <john> I'm hoping that will have something going for 0.2, but we've just begun doing preliminary brainstorming meetings
[13:58] <john> obviously to succeed at this goal, will need to nail down functionality very quickly
[13:58] <jonathan> and a guesstimate for 0.2 is July-August?
[13:58] <john> and, in the past this hasn't been our strength
[13:58] <Andy> more like September, I think
[13:59] amblin (~youwish@cosmo.itlab.musc.edu) left irc: Remote host closed the connection
[13:59] <mitchell> OK, last, last call for lingering quesitons
[13:59] <jonathan> thanks, that is helpful
[13:59] <mdubinko> john, if you'll be a round for a few minutes, I have a few questions..
[14:00] <john> sure, I'm happy to hang around to enter questions
[14:00] amblin (~youwish@cosmo.itlab.musc.edu) joined #chandler.
[14:00] <john> opps, should have been"answer", not enter
[14:00] <mitchell> OK, well we'll end the official session now.  
[14:00] Andi (~chatzill@63.90.156.153) left #chandler.