Portable DXR/meeting 102908

From CDOT Wiki
Jump to: navigation, search

<< Return to Portable DXR

Date  : Oct 29, 2008
Subject: Documenting the project
People : David Humphrey (humph), James Boston (jboston), Samer Ziadeh (samer)

13:34  * humph flips to pdxr
13:34 <@humph> where's jpau?
13:37 < jboston> I'm not sure.
13:37 < jboston> You're probably all wondering why I've called you here today....
13:38  * humph has a lot to do, would love to stay on target with time
13:39 < samer> yes, why are we here
13:39 < jboston> Two issues: 1) We haven't really done much and 2) we aren't sure about what our
                 'deliverables' are.
13:39 <@humph> how can I help?
13:39 < jboston> MLM wants us to meet with you regularly.
13:39 <@humph> ok, done.  now what can I do to help with 1) and 2)?
13:40 < jboston> I know this would go better if I could point you at the wiki to critique it, but
                 there hasn't been much activity.
13:40 <@humph> what have you guys done?
13:41 < samer> we had a meeting a couple of weeks ago, and we came out with a rough
               diagram/architecture
13:41 <@humph> you don't mean what I drew on the white board, right?
13:41 < samer> http://zenit.senecac.on.ca/wiki/index.php/Image:16-10-08_1801.jpg
13:41 < samer> no
13:41 < jboston> We had a meeting. We have some prj planning stuff to do. ie. Work Break Down
                 stuctures in ms project. But I don't think we need to worry abou that here. I was
                 thinking that perhaps we could produce some mockups to show you.
13:41 <@humph> guys, I need 5 mins, sorry. brb
13:41 < jboston> ok
13:42 < samer> ok
13:42 < samer> yeah she said we don't necessarly need ms project
13:42 < samer> just like a planning thing on the wiki or something
13:45 < jboston> MLM is being very kind.
13:45 < samer> we've been slacking for 6 weeks let's work our asses off for the next 6 weeks
13:45 < samer> if we meet once a week with humph, once a week as a group, and once a week with MLM
               we'll do good
13:45 < jboston> I think I can produce some mockups of how the updates will be presented to the end
                 user. Perhaps this will help her understand how the project works. I don't think
                 we've really made it clear to her.
13:46 < samer> yeah we've been really vague so far
13:46 < samer> like she sees potential in the development of the project, but the planning part is
               what worries everyone
13:47 < jboston> I'm still not sure, are we supposed to be planning the project or planning how to
                 plan to the project?
13:47 < samer> so far we've been planning how to plan the project
13:47 < samer> we need to plan the project
13:49 < jboston> But what are the deliverables for this semester? Is is the plan for planning? Or
                 the plan of the implementation?
13:50 < samer> for us it's different
13:50 < samer> we just need to fill out the wiki on a regular bases
13:50 < samer> that's how she's marking us
13:51 -!- jPau [jPau@B21045B7.6170312.E7B8CEC1.IP] has joined #PDXR
13:51 < jboston> Ah. So it should be a detailed description of the project in it's ideal form.
13:51 <@humph> back, sorry
13:51 < samer> it should be everything about the project
13:51  * humph reads scrollback
13:52 < samer> we need a nice description of the project on the wiki
13:52 < jboston> MLM did say that.
13:52 < samer> yeah
13:52 < samer> in our last meeting
13:52 < samer> jPau is here
13:52 <@humph> ok
13:52 <@humph> so you need to plan how this thing will work
13:52 < samer> yeah
13:52 <@humph> that can be done
13:53 <@humph> as I see it, you have a number of parts
13:53 < jboston> I think that's what we do. But then again, MLM has said that we are not doing
                implementation yet.
13:53 < samer> humph: also this http://zenit.senecac.on.ca/wiki/index.php/Image:16-10-08_1751.jpg
13:53 <@humph> jboston: yeah, not impl
13:53 <@humph> but *what* it will do
13:53 < samer> jboston: yeah just planning for the implmentation
13:53 < jboston> A platonic version.
13:53 <@humph> like you should be able to write in some detail what it is
13:53 <@humph> sure, shadows on the wall, and all that
13:54 < jPau> Hmm. Did Mary say that we need more business rules?
13:54 <@humph> I think you can do this quite easily
13:54 <@humph> let's quickly scope this beast
13:54 < samer> yes
13:54 < jboston> get your stuff ---> extension ---> push updates
13:54 <@humph> that diagram is missing a big piece
13:55 <@humph> yeah, the ext
13:55 < samer> humph: isn't that the Mozilla part of the diagram?
13:55 < jboston> I think ctyler is right. We should right our own server.
13:55 < jboston> s/right/write
13:56 <@humph> jboston: no
13:56 <@humph> that's crazy talk
13:56 <@humph> there are probably 20 single file C web servers you can embed
13:56 < jPau> If we have our own server, wouldn't we have to pay for it?
13:56 < jboston> That's what I thought. 
13:56 <@humph> here's what I see:
13:56 < jPau> That would go into the cost.. the stuff we learned in class
13:56 < samer> I think the server should be left till the end
13:56 < samer> we'll have the application and everything and then we'll see which server best fits it
13:57 <@humph> 1) Someone has to figure out how to mash a web server into an extensino.  There is
               research to be done there on all the options, compare them (which platform, license,
               complexity, etc.)
13:57 <@humph> samer: no
13:57  * humph shakes his head
13:57 < samer> oh
13:57 < samer> how come
13:57 <@humph> one of you could spend a week or so doing this research.  there is a ton written on it
13:57 < jboston> Ok. 
13:58 <@humph> because impl will mean taking whichever one you choose and doing the work
13:58 <@humph> design is figuring out which one to use
13:58 < samer> ah
13:58 <@humph> 2) you have to figure out how to a) get updates from my stuff; b) push updates to the
               ext
13:58 <@humph> this is somewhat non-trivial.  a) is a known entity.  but how do you do b)?  What do
               you push, how often, etc.
13:59 <@humph> you should be able to answer all those questions now
13:59 < jPau> We can do the research in the meeting this week..
13:59 <@humph> one of you needs to learn how to use my stuff so you can duplicate what I do
13:59 < jboston> Some kind of binary diff of a sqlite file maybe?
13:59 <@humph> this is the question
13:59 <@humph> it's a mixture of sqlite (binary diff, maybe) and source files
14:00 <@humph> how often do you update?  are their prefs for this?
14:00 <@humph> what other prefs are there?
14:00 <@humph> how do you deal with different branches, trees?
14:00 <@humph> do you support thunderbird code?
14:00 <@humph> etc.
14:00 < jboston> You plan to always have a version of dxr online?
14:00 <@humph> so many questions you can ask/answer
14:00 <@humph> that's not really relevant
14:00 <@humph> your stuff will be totally separate
14:01 < jboston> I ask because I wonder if our pdxr is based on snapshots of what you are doing, so
                to speak. But maybe I'm thinkin about it wrong.
14:01 <@humph> there is overlap, but let's now worry about that now
14:01 <@humph> my online version will likly update 24x7
14:02 <@humph> you might want to reserach what the delta size is between two versions of the
               backend, so you know how much needs to get pushed, and therefore how often
14:02 <@humph> extensions have auto-updating code built-in, so if you can package the content of the
               local web app within the extension, you can likely take advanatage of that
14:02 < jboston> Ok. I see the updating the snapshot available for the user as harder than pushing
                 the update. We'll need to figure out how you generate the database then?
14:02 <@humph> right
14:02 < jboston> aha
14:02 <@humph> I can show you that pretty quickly
14:03 < samer> is the DB sqlite?
14:03 <@humph> you can build your own dxr with an afternoon's work
14:03 <@humph> yes
14:03 < samer> I'd like to build it
14:04 <@humph> https://bugzilla.mozilla.org/show_bug.cgi?id=435814 and
14:04 <@humph> https://developer.mozilla.org/en/Dehydra_GCC
14:05 < jboston> Is bsmedberg's version available to the public? Not that yours isn't awesome. I'm
                 just curious.
14:05 <@humph> http://hg.mozilla.org/users/bsmedberg_mozilla.com/dxrpy/file/05453eb3a58a
14:05 <@humph> http://office.smedbergs.us:8080/file?p=xpcom%2Fglue%2FnsINIParser.cpp#l58
14:06 < jboston> Should I redact that link from public notes?
14:06 <@humph> probably
14:06 <@humph> the hg one is fine
14:06 < jboston> ok
14:07 <@humph> but don't get lost in that
14:07 <@humph> in some ways, your code is web app agnostic
14:07 <@humph> you're doing the system that allows his or mine to work
14:07 <@humph> another issue:
14:07 <@humph> you need to get a perl (and python if you want his stuff) interpreter into your ext
14:08 <@humph> or into an installer so it lives locally as well
14:08 <@humph> someone needs to research the best way to achieve this
14:09 <@humph> in short, there is no end to the planning you can do right now without writing any
               code
14:09 < jPau> ok
14:09 <@humph> one of you can reserch and write about how the ext update mechanism works
14:10 < jboston> Ok.
14:11 < jboston> I think we have concluded this meeting?
14:11 <@humph> so anything else today?
14:12 <@humph> yeah
14:12 <@humph> I have some links, actually
14:12 <@humph> sec
14:12 < jboston> ok
14:12 <@humph> http://en.wikipedia.org/wiki/Tiny_web_servers
14:13 <@humph> I found that quite helpful when I was looking at this problem
14:13 < jPau> Thanks. That's really helpful.. we had trouble finding a good web server.
14:14 <@humph> there's also a good ibm article if you go googling
14:14 < jboston> i think I have a link to that
14:14 <@humph> so in short, get blogging, get researching, and write about what you find.  start
               asking each other questions (how will it do this? do we have a pref for that?) and
               document your answers
14:14 <@humph> before long it will be designed.
14:16 <@humph> ok, I'll be in #seneca as always, but idle here if you want to ping me.
14:16 < jboston> ok