**** BEGIN LOGGING AT Thu Jan 24 02:59:57 2008 Jan 24 04:48:43 alright, back to coding Jan 24 05:37:37 hi Jan 24 09:54:03 mornin Jan 24 09:54:21 morning Jan 24 13:47:53 dammit Jan 24 13:47:59 it seems this munich thing is already sold out Jan 24 13:48:18 you mean the android tour? Jan 24 13:48:40 yeah Jan 24 13:49:06 yeah Jan 24 13:49:14 i already noticed that Jan 24 13:49:16 :( Jan 24 13:49:19 that's pretty lame Jan 24 13:49:32 yep Jan 24 13:55:51 I wrote an email to the dude in Germany, and Mr. Morrill Jan 24 13:56:03 Morril Jan 24 14:00:08 cool Jan 24 14:04:15 there is a munich tour?! Jan 24 14:06:25 they should open a second then Jan 24 14:33:28 kristian-m, cutmasta, write them yourselves... the more people that do, the more they're going to be under pressure to meet the demand Jan 24 14:33:37 (sorry dan:-) Jan 24 15:41:00 im querying a cursor but I only want it to return the first row... Jan 24 15:41:25 how am I supposed to do that? Jan 24 15:50:17 are these things supposed to have motion sensors? Jan 24 20:11:14 hi all Jan 24 20:29:13 good evening Jan 24 20:29:20 howdy Jan 24 21:25:37 yawn Jan 24 21:28:00 yawn Jan 24 21:28:26 guess we need one of those loaded new coke soft drinks Jan 24 21:33:00 heh Jan 24 21:36:39 who came to the Campfire yesterday? Jan 24 21:36:43 And what did you think? Jan 24 21:37:04 morrildl: What did you guys do during the Campfire? Jan 24 21:37:14 you didn't go? Jan 24 21:37:42 jasta: Unfortunately I was too busy fixing bugs until late :) Jan 24 21:38:40 romainguy: mostly we just talked a lot :) Jan 24 21:38:47 :) Jan 24 21:39:24 * davidw wonders if a 'campfire' in the middle of silicon valley wouldn't feel a bit.... well... corny Jan 24 21:39:38 that's what i said :) Jan 24 21:39:48 you guys should've rented out a bar ;) Jan 24 21:40:54 jasta: well, there WAS beer Jan 24 21:41:14 morrildl: hows the NIO issue coming along? Jan 24 21:41:16 davidw: the first one we did was cornier than this one ;) Jan 24 21:41:28 jerkface03: should be fixed in the next SDK we release Jan 24 21:41:41 jerkface03, what's the NIO issue? Jan 24 21:41:42 morrildl: will the next SDK release be available before the challenge ends? Jan 24 21:41:53 davidw: NIO not implemented :) Jan 24 21:42:04 jerkface03: more specifically, code for a whole bunch of additional java.* packages got pulled in Jan 24 21:42:15 * davidw is wondering about the radio button thing... romainguy fixed it in like 5 minutes, but it wasn't in the latest release Jan 24 21:42:53 davidw: The fix did not happen in the M3 branch Jan 24 21:45:17 * jasta is getting excited now that the round 1 deadline is approaching ;) Jan 24 21:45:52 i've had to waste a ton of time on my project working on client/server crap. i hope that i'm judged fairly if my UI sucks ;) Jan 24 21:46:03 sif i havent' started Jan 24 21:46:11 i should start shouldn't i Jan 24 21:46:21 well, uhh, yeah :) Jan 24 21:46:37 when's the round one deadline? Jan 24 21:46:42 although my project was ambitious enough that i needed to start in dec. Jan 24 21:46:43 march 4 Jan 24 21:46:44 or something Jan 24 21:47:01 mines insanely ambitious Jan 24 21:47:06 i have a year to do it at uni Jan 24 21:47:08 that should be fun then Jan 24 21:47:13 1 month for compo Jan 24 21:48:41 mine will require a robust client/server architecture, where i must implement both :( Jan 24 21:49:07 although, i'm reusing a lot of existing code (both libraries and code i've written for previous projects) Jan 24 21:49:46 it's fortunate that giftd's meta data extraction and share database code is sufficiently generalized ;) Jan 24 21:51:31 morrildl, romainguy what happens to AndroidManifest.xml in the 'compiled' version? Jan 24 21:52:17 it is converted to some butchered WBXML. Jan 24 21:52:41 by butchered, i mean, it does not seem to comply with the WBXML specification, but it very much so seems to be. Jan 24 21:52:58 I want to hack it without recompiling it Jan 24 21:55:58 well, good luck. benno has some documentation on the format, i believe. Jan 24 21:59:19 hrm...looks like you'd need to fiddle with a number of things Jan 24 21:59:23 pass, for the time being, I guess Jan 24 23:35:37 davidw: as jasta said, the various .xml files are "pre-parsed" into a binary format for fast reads at runtime Jan 24 23:36:07 I don't know anything about WBXML, but I suspect any similarity is coincidental Jan 24 23:36:30 morrildl, hopefully at some point in the future, that machinery will be exposed... Jan 24 23:36:51 davidw: well, it will be when the source drops :) Jan 24 23:36:57 what do you mean by machinery? Jan 24 23:37:04 are you saying you want to fuss around with it? Jan 24 23:37:50 im guessing he wants to make real time changes to his app? Jan 24 23:39:32 morrildl, what I want to do is make it so that generating new Hecl apps is very easy. For instance, in the Java ME version, I have a tool, aptly called "JarHack" that can fiddle with the name of the thing and create a new .jar Jan 24 23:40:03 you don't even have to have Sun's WTK around to do it Jan 24 23:40:25 I'd like to be able to do the same thing with Hecl and Android. Replacing the res/raw/script.hcl file is easy enough with zip Jan 24 23:41:47 hmm Jan 24 23:42:42 so then you're saying that you want to "synthesize" or construct a .apk file from within the device for a specific Hecl script, and then write it to storage, so that the script appears to be its own app and gets its own icon, etc? Jan 24 23:42:49 it's not my most pressing problem at the moment, but it's a 'nice touch' that needs to be there sooner or later Jan 24 23:43:41 morrildl: I agree, it looks like the format deviates from WBXML sufficiently enough that it seems to have just been reinvented. Jan 24 23:43:45 morrildl, from within the device is frosting on the cake - I'd be happy to start with changing an existing apk outside the device, without having to open it up and use the provided tools Jan 24 23:43:59 Probably to get away from the clutches of the OMA :) Jan 24 23:44:05 davidw: hmm, so you want to munge .apk files. Jan 24 23:44:11 from within the device would be quite cool though Jan 24 23:44:14 morrildl, yeah Jan 24 23:44:26 jasta: probably just because it was a particularly efficient way to do it Jan 24 23:44:35 or at least hte most efficient we could think of :) Jan 24 23:44:36 like I said, it's not an immediate, big problem, but it's on the radar Jan 24 23:44:41 david Jan 24 23:44:47 davidw: hmm, okay Jan 24 23:44:54 it would help if yo uentered that as an issue in the Issue Tracker Jan 24 23:44:58 sure Jan 24 23:45:02 then I can get into the internal system and harass people about it :) Jan 24 23:46:32 ok, cool, thanks Jan 24 23:46:51 morrildl: WBXML was designed to be efficient in terms of size. One of the advantages is that applications can describe a string table in their specification. Jan 24 23:47:14 so that it is XML, but the transport contains no string XML tag literals. Jan 24 23:47:14 jasta: android's format is designed to be efficient to read Jan 24 23:47:31 morrildl: Yes, obviously ;) Jan 24 23:47:36 :) Jan 24 23:47:43 I'm just explaining what WBXML was, since you did not know and I feel like you should :) Jan 24 23:47:47 ahh Jan 24 23:47:49 okay then :) Jan 24 23:47:54 * morrildl is appreciative Jan 24 23:47:57 hehe Jan 24 23:48:12 I am using WBXML (with SyncML) in my ADC application. Jan 24 23:48:15 what is the format? Jan 24 23:48:25 and i have, unfortunately, had to implement most of it myself. Jan 24 23:48:43 the state of SyncML client libraries (in any language) is a sad state of affairs. Jan 24 23:48:46 meaning, I assume it's not just <0x2a 0x42="foo"/> Jan 24 23:49:21 is it like zero-delimited or something? Jan 24 23:49:27 or is it mixed ascii and binary or something? Jan 24 23:49:41 morrildl: Well it's not text, but that is in the spirit of how it was designed. There is/can be a string table preceding the document, which maps integer values to the tags and attributes. Jan 24 23:50:11 morrildl: It's mostly all size-delimited, IIRC. Maybe it's NUL, though. Jan 24 23:50:45 ahh Jan 24 23:50:50 The cool thing, like I said, is that specific applications defining documents in WBXML can, as part of their specification, include the literal integers for the string table so that it need not be communicated. Jan 24 23:51:13 you mean in the DTD? Jan 24 23:51:30 Not precisely, but yes basically. Jan 24 23:51:52 cool Jan 24 23:51:55 uh oh... Jan 24 23:52:19 - it looks at the uniqueness of the class:-/ Jan 24 23:52:52 wbxml was defined by OMA, though, and is backed by the big guys. Motorola, Nokia, etc. Jan 24 23:54:26 well that puts a damper on things Jan 24 23:55:19 morrildl: The submission guidelines for the ADC seem to suggest you only need to provide an APK, is that true? Source code is not necessary? Jan 24 23:55:27 jasta: correct Jan 24 23:55:35 of course your app has to run against one of the public SDKs Jan 24 23:56:25 Hmm, I have a server component which is written in C, that is very much so a part of my submission. I would prefer to distribute that with only source, to be compiled by the judge. Jan 24 23:56:29 Is that sort of a no-no? Jan 24 23:57:04 jasta, why not just put it on a public server? Jan 24 23:57:31 Well, because that would be a copyright violation in the United States :) Jan 24 23:57:49 I'd prefer my application not be illegal for Google to use. Jan 24 23:58:25 ? Jan 24 23:58:35 gpl'd? Jan 24 23:59:16 My application is a media synchronization library and accompanying music player. It is designed to let you synchronize meta data and stream content from a hosting server, e.g. your home computer. Jan 24 23:59:35 find some public domain music to put on it to start people out with Jan 24 23:59:41 So, if I provided my server, and my content, in order to judge the application they would need to copy the music I have on my machine. Jan 25 00:00:07 it's not like you have to open it to everyone right at the start Jan 25 00:00:08 I honestly doubt I could find enough public domain music to test all the features accurately. Jan 25 00:00:18 I bet you could Jan 25 00:00:22 But perhaps. Jan 25 00:00:28 might not be *good* :-) Jan 25 00:00:48 Gotta be more than like 10 songs, obviously. :) Jan 25 00:01:07 Also, I have last.fm integration in my music player, and I bet last.fm has got no clue about a bunch of public domain shit ;) Jan 25 00:02:14 http://www.pdinfo.com/ Jan 25 00:02:34 but I see your point, it would be better if I could control the initial experience better. Jan 25 00:04:27 I think maybe I'll provide both. A sample server, and the source code to the real server. Jan 25 00:04:59 jasta: you could also just record yourself saying "this is a test" several times, and then fake the ID3 data Jan 25 00:05:26 that would also have the advantage of keeping the files very small :) Jan 25 00:05:48 11KHz mono, 96kbps sampling rate :) Jan 25 00:09:02 morrildl, I filed a new issue 85, but in checking how things work, I discovered that class uniqueness determines app uniqueness, which puts a bit of a kibosh on things in any event, so that got its own mailing list post Jan 25 00:14:06 morrildl: but that hardly tests my program :) Jan 25 00:15:23 actually, i have a great idea. Jan 25 00:15:49 I can just copy my personal music collection and rewrite every file's audio stream as a recording Jan 25 00:16:06 which is actually exactly what morrildl said, I just realized. Jan 25 00:16:25 yeah, that's what i'll do. Jan 25 00:16:43 i actually wonder if my program violates some new insane DMCA provision. Jan 25 00:17:36 heh Jan 25 00:17:59 * morrildl has opinions, but probably nothing he can say :P Jan 25 00:21:30 i can't fathom this being legally different than burning a cd or copying music to an ipod or similar device? Jan 25 00:21:38 i mean, ...right? :) Jan 25 00:24:11 That is the assumption I have been going on for my own personal CD collection Jan 25 00:26:50 I think those record companies are completely coming unglued, but that's straying pretty far from the topic at hand... Jan 25 00:33:03 sleep time for me... thanks for the help morrildl, later all Jan 25 00:33:10 davidw: np Jan 25 00:34:23 jasta: I'd suggest putting up a site where you include the server source (or at least a binary), one of the junk MP3 files, and a script to clone it and munge the clones' ID3 data, if you really want to Jan 25 00:34:38 however it's worth noting that not all the judges are likely to bother Jan 25 00:34:48 so you'll want to have "something" up and running, regardless Jan 25 00:35:09 (it should be perfectly okay for that "something" to be a local-only mock) Jan 25 00:41:17 morrildl: Hmm. Jan 25 00:41:36 I guess I just don't understand how a "local only" set up could be used for judging? Jan 25 00:41:44 It's not how anyone will ever use my application? Jan 25 00:42:34 I will provide the server source and instructions on how to set it up, as well as a publicly hosted server running my software that the client can be connected to. Jan 25 00:43:17 So lazy judges can connect to a preinstalled server. Jan 25 00:43:25 But I'd hate to think that will be most or all of them. Jan 25 00:43:51 I suppose if they like my idea they will try harder to make it work right :) Jan 25 00:45:11 jasta: I assume you have or can create a provider interface, like GetMp3Info that has methods like getListOfMp3s() and playMp3() Jan 25 00:45:17 the "real" version talks to the network and stuff Jan 25 00:46:06 the "fake" version would have a single MP3 stored in your resources, and have an implementation that always returned a hard-coded fake list of MP3s Jan 25 00:46:20 of course that's probably a decent amount of work for little gain Jan 25 00:46:27 and it doesn't show off your actual sync component Jan 25 00:46:41 so I guess it comes down to how reliable your hosting is Jan 25 00:51:12 If my server is down, I cannot require that the user set the server up themselves? Can ADC participants not assume that judges must follow installation instructions provided explicitly by the submission? Jan 25 00:53:34 I thought once about making the sync function have a dummy mode which injects random useless data but it actually seemed like too much work considering the server is not so complex. Jan 25 00:53:57 and that would merely be developer testing, no end user would ever care about that. Jan 25 00:54:56 Also, the amount of data required for a meaningful experience is not insignificant. Sure, I could have a single real MP3 that is just nonsense loop over-and-over, but remember: meta data is the important piece. Jan 25 00:55:04 and there application is designed to accept LOTS of it. Jan 25 00:55:09 s/there/the/ Jan 25 00:55:09 jasta meant: and the application is designed to accept LOTS of it. Jan 25 00:55:27 * jasta raises an eyebrow at infobot Jan 25 00:55:32 jasta: can't you have the same real mp3 over and over again with different metadata? Jan 25 00:56:08 romainguy: well I think a major part of jasta's is the actual media-sync system (and that's the part I am most interested in) Jan 25 00:56:46 I can easily envision a background-sync process that magically keeps the media content on the SD card "current", while deferring to say the standard Media Player UI for actual playback and stuff Jan 25 00:57:02 jasta: am I close? Jan 25 00:57:57 morrildl: You are dead-on. The media player UI uses a content provider (from the sync process) which has all the meta data ready. The media player UI then makes a choice, which prompts the service to either serve the content from cache, or go get it. Jan 25 00:58:08 There is a strong separation between the two pieces. Jan 25 00:58:30 The UI does nothing but deal with meta data from the content provider and ask the service to make content available for playback. Jan 25 01:00:50 It would seem a little hollow to mock out the guts of that ContentProvider Jan 25 01:01:10 It just seems awkward to inject a ton of dummy data into the content provider. It doesn't even "test" anything. Jan 25 01:01:32 In that case, I guess I'd suggest setting things up so that you can fetch the metadata and stuff from a static server Jan 25 01:01:35 I mean, I can see it being useful to test the server-side putting a bunch of dummy data in to test various features of the client. Jan 25 01:01:42 I think Google Pages lets you upload arbitrary files Jan 25 01:01:55 It'd all be static, but it would be reliablyup Jan 25 01:02:46 Perhaps. I'm not that worried about the reliability of my hosting, though. Jan 25 01:03:04 I just would feel better if the judges had a basic willingness to make the software work as it was intended to :) Jan 25 01:04:30 The server is currently as simple as just saying ./syncml-server --music ~/Music. It spits out a bunch of "scanning" messages and finally "ready to serve." It binds a single TCP port, easily forwarded. Jan 25 01:04:55 And the client is configurable to add or remove various server sources (consisting of hostname and port). Jan 25 01:05:35 i've concentrated most of my efforts in this area at present. i haven't even started the music player yet. Jan 25 01:05:56 I figure, I can take a crack at a basic media player taking advantage of all this meta data, and wait for someone to make a less sucky one in the years to come that I can attach my service to ;) Jan 25 01:08:03 jasta: well, let me put it like this Jan 25 01:08:10 *I* would have no problem following your directions Jan 25 01:08:27 But I can't speak for the judges from all the other Alliance member companies Jan 25 01:08:37 (and I don't know that I myself will actually be a judge) Jan 25 01:08:47 obviously my end goal is to make this easy for users to set up, but it will not fundamentally be different than installing a server, and then configuring the client. Jan 25 01:09:05 it can't ever be simpler than that :) Jan 25 01:09:13 jasta: indeed :) Jan 25 01:09:36 brb, heading home Jan 25 01:09:38 heh, maybe instructions of "install this software and point it at your own CD collection" will provide the required motivation :) Jan 25 01:10:09 I'll try to provide both, just in case. A static server I host (I have stable hosting I can exploit for this), as well as fully instructions for setting up the server. Jan 25 01:10:21 But the first round of the challenge (this one) will only include Linux server support. Jan 25 01:10:30 I'd much prefer some other goon make a Windows application :) Jan 25 01:10:46 and i don't want to deal with all the DRM hell that exists on Windows environments. Jan 25 01:10:58 anyway, bbl. Jan 25 02:10:09 can I define a color class in an xml file? **** ENDING LOGGING AT Fri Jan 25 02:59:57 2008