**** BEGIN LOGGING AT Thu Aug 14 02:59:57 2008 Aug 14 08:46:09 romainguy: are you here? Aug 14 08:46:31 Now the "Spiegel" (one of the biggest magazines in germany) takes up the septembre rumours :D Aug 14 08:52:38 anno^da_: and as expected it's bad: "spiegelt das Filmchen nur die zunehmende Verstimmung innerhalb der Open Handset Alliance wider." Aug 14 08:54:49 Yeah I know. :) Aug 14 08:55:01 But that false rumours could be stopped. ;) Aug 14 08:56:39 spe is too close Aug 14 08:56:41 sep Aug 14 08:59:45 It is but the rumours will go on Aug 14 09:00:10 Google News, the last 3 news regarding Android pick up the sep rumours :D Aug 14 09:01:00 It is funny to follow the copy and paste journalism Aug 14 09:03:58 http://www.ne-na.de/A556D3/NENA/NENA_NEU.nsf/0/EABD81A289926C9BC125749F003C62B5?OpenDocument <- lol that came out on the 8th of august for all the germans here. Aug 14 09:04:36 "„Google wird ein Endgerät auf den Markt bringen, das als persönlicher Informations- und Kommunikationsmanager fungiert." <- hihi Aug 14 09:05:01 I'm loving the analysts :P Aug 14 09:06:17 alex2308: And did you have it in your hands already? Aug 14 09:07:27 he is working for the Telekom not TMo Aug 14 09:08:05 Ok - too bad Aug 14 09:08:57 And even if he had seen it, he had to sign a NDA :D Aug 14 09:09:48 Well you wont see much at the Telekom Aug 14 09:10:09 But at TMo you might see someting Aug 14 09:10:22 Yeah at TMo you can see sth. Aug 14 09:10:29 :) Aug 14 09:10:56 It will be my first time with TMo... :D Aug 14 09:11:00 depends where if he works in Darmstadt or Bonn. In Darmstadt it will be very unlikely for him seeing the devuce. Aug 14 09:11:34 Hmm well I will buy just the device and keep waiting for more interesting plans. Aug 14 09:11:50 That might be an option too Aug 14 09:12:19 I think that it takes perhaps 1 year and we will see the data flatrates at congstar and the other prepaid providers. Aug 14 09:12:34 (ok Congstar == TMo :D) Aug 14 09:12:56 Didn't I mention moobiair? T-Mobile reseller... Aug 14 09:13:06 ah yeah you mentioned that Aug 14 09:13:40 is there a 2 years contract ? Aug 14 09:13:45 No Aug 14 09:14:18 ok 6 months right ? Aug 14 09:14:23 Seems so Aug 14 09:14:35 spykid ? Aug 14 09:14:47 err, yes i signed a NDA Aug 14 09:15:05 oct is what everyone says Aug 14 09:15:44 anno^da_ i work in darmstadt, but the division im in is so high inside the enterprise structure that i have access to everything Aug 14 09:15:58 my id card even lets me enter every building Aug 14 09:16:22 well ok :) Aug 14 09:16:33 i can even enter deutsche post buildings ^^ Aug 14 09:16:51 me too, at least some of them ,-) Aug 14 09:17:07 the public ones? ;) Aug 14 09:18:15 exactly ;) Aug 14 09:18:35 do they have android phones :) Aug 14 09:19:31 Ah thats why alex2308 is so silent in the last time Aug 14 09:19:33 ;) Aug 14 09:20:00 ;) Aug 14 09:20:52 no, its just a lot of work Aug 14 09:20:58 and im looking for a new job atm Aug 14 09:22:18 hmm, im bind to a service in activity.oncreate() and the service calls a callback method of the activity right on binding. i get "activity not running" errors when i try to do any window/dialog stuff in the callback. Aug 14 09:23:00 so where is the best place to bind to the service. doc says "do not use" onPostCreate() though i think that might work in my case? Aug 14 09:24:26 onresume? Aug 14 09:25:21 strange I'm binding as well in onCreate Aug 14 09:27:44 i think its a bit buggy there Aug 14 09:28:16 when i tried registering services i realized different behaviour even when not changing my code Aug 14 09:28:52 anno: i had not problems until i called the callback method right after registering the callback right after binding? Aug 14 09:30:12 alex: yeah, i got that error first a while back ago, but only rarely. but atm i get it almost every second binding. looks like some kind of race condition to me. Aug 14 09:45:19 spykid: "Sollte es stimmen, würde die Android-Plattform das Licht der Welt mehr als ein halbes Jahr früher als bisher gedacht erblicken." <- haha. Aug 14 09:45:50 Q4 08 - September = ? :P haha man man man Journalisten und rechnen :P Aug 14 09:46:53 (oh back to english :)) Aug 14 09:48:57 anno^da_: they were refering to the rumored delay to Q2 2009, that's how they got to 6 months ;) Aug 14 09:50:00 ah well ok :) Aug 14 09:50:17 They are taking rumours to build up new rumours :D Aug 14 09:50:36 A web 2.0 mashup. Aug 14 09:51:06 anno^da_: I think I should ask someone to use this headline: "Android delayed! Android teams laughs its ass off about rumors and now they are unable to work" :D Aug 14 09:52:20 Well give an interview to the magazines as a "ANDROID DEVELOPER" and you will be quoted ;) Aug 14 09:52:35 no matter if it is right or wrong :) Aug 14 09:52:58 And if you made it in one magazin it gets spread all over the internet blogs and news sites Aug 14 09:53:39 Na, that's not right! Your words have to fit in their agenda ;) Aug 14 09:54:00 Ok :D. Then you have to play with the latest rumours. :) Aug 14 09:54:10 Yep Aug 14 09:54:22 So take in consideration that it is delayed and now it could be september. Aug 14 09:54:50 So lets create another rumour. :) Aug 14 09:55:43 and we have very trusted sources. Aug 14 09:56:23 Isn't alex2308 very trustworthy :D Aug 14 09:57:19 Well yeah I would say he is. Aug 14 09:57:50 So alex2308 tell us something new about the android system :) Aug 14 10:00:03 And lets add some blurry videos from a HTC device. Perhaps we take the same video and just cut the scenes in a different order and change the colors a little bit. Aug 14 10:00:19 anno^da_: He has signed a NDA and he may not even be allowed to tell us how much he has to pay if he discloses something :D Aug 14 10:01:21 is there an android appstore being setup? Aug 14 10:01:35 there will be some store. Aug 14 10:01:56 But rumours could tell you sth different muthu Aug 14 10:01:58 ;) Aug 14 10:02:13 haha Aug 14 10:02:34 hope android comes out soon to break those rumors ;) Aug 14 10:04:47 anno^da_: Augsburg? Aug 14 10:05:47 not really :) Aug 14 10:06:20 anno^da_: At the moment? Aug 14 10:06:30 Or Aachen? Aug 14 10:06:51 Not Aachen and not Augsburg :P Aug 14 10:07:15 From my IP adress you wont see sth. :) Aug 14 10:10:02 Ok Aug 14 10:17:18 muthu tmo has some kind of appstore, for android, too ofc Aug 14 10:17:38 looks like it Aug 14 10:17:50 cant tell more :p Aug 14 10:18:10 tmo appstore is all in the news Aug 14 10:18:18 very exciting Aug 14 10:19:19 haha alex2308 said: "An app store will be available in September by T-Mobile" - and now spread it :P Aug 14 10:19:51 http://users.skynet.be/fa018426/Georgia.jpg Aug 14 10:20:29 lol Aug 14 10:20:37 lol Aug 14 10:21:01 haha, love the first response :) Aug 14 10:33:36 anyone doing MS updates? Aug 14 10:33:47 download is extremely slow Aug 14 10:33:53 +here Aug 14 12:11:53 can we populate ContextMenu on MapViews too? (ListView works fine ;/ ) Aug 14 12:13:22 hey plusminus_ Aug 14 12:13:31 hi Aug 14 12:13:39 mapviews already have this zoom thing as the contextmenu Aug 14 12:14:08 ok .. then how to override it with the built in ContextMenu? Aug 14 12:14:28 until now I launched a SubActivity in: Aug 14 12:14:36 never tried it Aug 14 12:14:42 setOnLongPressListener(new OnLongPressListener() { public boolean onLongPress(...) { Aug 14 12:14:48 / here Aug 14 12:15:07 which worked fine, but I decided to do it with system-internals Aug 14 12:16:57 do you have an overlay? Aug 14 12:17:04 yes Aug 14 12:17:17 maybe you have to add the contextmenu to the overlay Aug 14 12:31:53 contextmenus can be applied only to views not overlay Aug 14 12:32:15 check your prvmsg :-) Aug 14 12:44:12 global thermonuclear war Aug 14 12:44:14 ops Aug 14 12:51:27 run run run Aug 14 13:02:35 cool, i got a black htc dream Aug 14 13:10:12 alex2308: You are breaking your NDA :D Aug 14 13:12:01 maybe i was jk to see what happens ^^ Aug 14 13:12:48 And btw I'm sure it's name won't be 'Dream' ;) Aug 14 13:13:05 why not? Aug 14 13:13:22 everybody knows it. I would stay at "Dream" Aug 14 13:13:24 but maybe they call it "minusplus" Aug 14 13:13:26 ... Aug 14 13:13:43 Doppeltplusgut Aug 14 13:13:54 or "Trey's Bad Ass Android Phone" Aug 14 13:14:16 "ePenis" Aug 14 13:14:32 as a competitor for the "iPhone" ePenis Aug 14 13:14:46 or "HamClub" because it will be big as a baby... Aug 14 13:14:51 for when you don't have to compensate for anything... Aug 14 13:15:09 true, its really big Aug 14 13:16:58 It won't be called 'Dream' because it would be too easy to spot it on FCC Aug 14 13:18:01 maybe its listed delayed Aug 14 13:19:09 If there doesn't appear a 'Dream' in the next 4 weeks at the FCC it won't be called 'Dream' ;) Aug 14 13:20:22 why not search for htc? Aug 14 13:21:43 aw crap i want a contexmenu on a MapView o_O Aug 14 13:23:56 aw crap, i want more money $_s Aug 14 13:32:32 what are popular eventhandlers? where are typical hooks? Aug 14 13:32:38 (windows environment) Aug 14 13:33:07 i know keyboard and mouse, but im not sure why apps hook them (except keyloggers) Aug 14 14:57:09 what info is public on the speech module? i see a speech class? is it for speech recognition? Aug 14 14:58:56 http://www.youtube.com/watch?v=FWeD5KXx5WI Aug 14 14:59:33 1:10 Aug 14 15:13:10 re Aug 14 15:30:48 yawn Aug 14 15:35:07 gphone finally killing sms/mms Aug 14 15:35:20 huh? Aug 14 15:35:31 ignore him, he's just a troll Aug 14 15:35:36 just random thoughts Aug 14 15:38:47 with gtalk preinstalled on every phone Aug 14 15:38:52 (and gtalk is free) Aug 14 15:39:02 why would you send a sms/mms, when you could send a IM Aug 14 15:39:20 have you guys seen the break down of how much it costs to send sms messages? Aug 14 15:39:34 assuming you don't have an unlimited plan Aug 14 15:39:37 never sent one in my whole life Aug 14 15:39:58 you know the thing that keeps me from sending them? Aug 14 15:40:10 my craptastic samsung phone Aug 14 15:40:15 i think its 99cent per mms here Aug 14 15:40:45 it's something like 20 cents a message here, but they just started billing for incoming and outgoing Aug 14 15:41:08 so they charge 40 cent per message if you look at the sender and receivers fees. Aug 14 15:41:35 I read a break down on the cost of the bandwidth and it was ridiculous Aug 14 15:41:38 not fair, what if i dont want mms from you? you charge me then? Aug 14 15:43:25 http://gthing.net/the-true-price-of-sms-messages Aug 14 15:43:33 interesting read if you have a few minutes Aug 14 15:44:27 Costs of transferring 2,560 MP3s via my ISP: $1 via SMS: $61,356,851.20 Aug 14 15:44:32 we all know this Aug 14 15:44:36 we aint stupid Aug 14 15:44:46 (maybe except me) Aug 14 15:47:42 teehee, playing hooky today Aug 14 16:24:17 re Aug 14 16:26:25 wb Aug 14 16:28:21 morning romain Aug 14 16:29:04 romainguy: morning Aug 14 16:29:06 mornin jasta Aug 14 16:29:38 oi SanMehat Aug 14 16:32:32 *up* what info is public on the speech module? i see a speech class? is it for speech recognition? Aug 14 16:32:43 is there any voip application for android? Aug 14 16:44:04 re Aug 14 16:44:16 oh I love the instable osx sometimes :> Aug 14 16:44:46 may i quote this? my friend says osx is more stable than vista Aug 14 16:45:02 i managed to crash osx 3 times - and i dont even have a mac myself! Aug 14 16:48:27 well the mbp here has a uptime now for 24 days counted till the last updates Aug 14 16:48:48 but it is definetly not that stable than my ubuntu machine is Aug 14 16:50:33 But the point is that I love working with it. It just feels a lot better than Vista. But that's a personal view. And I hate the Vista layout and the windows app design. It is just ugly. But no bashing it is just my opinion and everyone should use the system that fits best his needs. Aug 14 16:51:23 anno^da_: i think that the reason folks enter this holy war is because more often than not, you are not permitted to use the system that best fits your needs. Aug 14 16:51:59 most corporate environments, for example, even when they do let you use Linux, it's some ancient version of RHEL. Unanimously hated by everyone, Windows, Mac, and Linux users alike :) Aug 14 16:53:25 anno^da_: thats just what i say - use whatever you like. current OSs are all very advanced Aug 14 16:53:37 you can use mac/linux/win/whatever here Aug 14 16:53:53 most use windows, i think mac is at 4% (which is high) Aug 14 16:54:47 market share is a silly thing to quote. it is clearly the overwhelming favorite of business and typical home users, but in certain communities you find staggering use of Linux and Mac by contrast. Aug 14 16:55:42 wouldnt say that is it "the clearly the overwhelming favorite of business and typical home users" Aug 14 16:55:45 For example, in the open source community, it is quite reliable to assume that you are dealing with someone who primarily works with -- or at least is familiar with -- Linux. Aug 14 16:55:59 me and my friends dislike osx and apple, it is clearly my least fav Aug 14 16:56:10 jasta true Aug 14 16:56:37 I was referring to Windows being the overwhelming favorite, obviously. Aug 14 16:56:54 ops, my fault, did get it wrong then Aug 14 16:57:03 But again, market share is an average over all segments, but that doesn't tell you anything you can actually use. Aug 14 16:57:20 Beacuse there are communities all over that have wildly different percentages they could quote. Aug 14 16:58:16 I find it rare, even, to interact with people who use Windows for development. In my world, OS X and Linux are the majority. Aug 14 16:59:04 windows and linux are in mine Aug 14 16:59:11 mac is really really rare Aug 14 16:59:29 "dislike osx and apple" ... and have never used one :) Aug 14 16:59:31 All I wanted to stress was that quoting market share doesn't tell you anything interesting. None of us are particularly concerned with what the typical home PC user does. Aug 14 16:59:31 only Macs in mine :) Aug 14 16:59:49 romainguy: !! Aug 14 16:59:50 :) Aug 14 17:00:12 but i know some guys who switched to mac (bc its trendy and ofc bc of the design) after a while they started emulating windows, then dual boot, then switched back to windows. but those were average users, no programmers/it-specialists Aug 14 17:00:46 yeah thats ok Aug 14 17:03:19 My observations have led me to the conclusion that as ability increases, so too does the distain for Windows. There is a breaking point for most that causes them to leave it for an alternative. Aug 14 17:03:23 ;) Aug 14 17:03:51 depends, i have a lot of very skilled friends who prefer windows Aug 14 17:03:59 skilled developers Aug 14 17:04:16 in security you go to linux/unix/bsd Aug 14 17:04:19 I'm sure you do. Aug 14 17:05:18 Skilled :) what does skilled mean ... It is all about what you do with the system. And what you prefer. Aug 14 17:05:45 uhm, everything? Aug 14 17:06:16 A lot people use windows because they got it installed on their first pc from the discounter. Learned it at school and had to use it at work. Aug 14 17:06:30 full ack Aug 14 17:07:11 They just dont know about any other system. For a normal user Winows == operating system. (the only one) Aug 14 17:07:31 and talking about documents == .doc Aug 14 17:07:50 and web == .htm Aug 14 17:07:52 :) Aug 14 17:07:54 :D Aug 14 17:08:11 and search == google Aug 14 17:08:27 well yeah Aug 14 17:08:28 for sure Aug 14 17:08:30 duh Aug 14 17:08:33 please do not use that pseudomathematical gibberish. Aug 14 17:08:44 dael with it Aug 14 17:08:48 /s/dael/deal Aug 14 17:09:02 if you think "is" means "equals", then i hate you :) Aug 14 17:09:11 haha jasta :) Aug 14 17:09:35 lol. Aug 14 17:09:42 I dont care about that even if you hate me for that. ;) Aug 14 17:09:47 have you guys already discussed the appleinsider story? Aug 14 17:09:53 try #apple Aug 14 17:10:04 it's an article on android Aug 14 17:10:08 http://www.appleinsider.com/articles/08/08/13/google_android_flaws_pushing_software_firms_towards_iphone.html Aug 14 17:10:10 werd. Aug 14 17:10:19 nc. Aug 14 17:10:33 ouch Aug 14 17:10:52 it must be true Aug 14 17:11:17 the author was reading my mind Aug 14 17:12:10 well, actually i agree with that article. Aug 14 17:12:15 for the most part Aug 14 17:13:26 whats the short version of the article? ^^ Aug 14 17:13:47 give us an updated skd Aug 14 17:13:49 sdk Aug 14 17:13:58 ^^ Aug 14 17:14:18 I think most of it is rubbish Aug 14 17:15:38 Though most of the quotes are accurate. Aug 14 17:17:47 i remain convinced that professional developers would be crazy to adopt android now. Aug 14 17:18:01 and they would have been crayz to do it months ago too. Aug 14 17:18:26 the platform has essentially been a secret from day one, despite there being two SDK releases. Aug 14 17:18:34 jasta: you want a pony? Aug 14 17:18:41 nm, i've got work to do Aug 14 17:18:42 * SanMehat & Aug 14 17:19:19 jasta: And it's perfectly normal, Android was introduced as a preview release Aug 14 17:19:41 we announced the first phones by the end of the year and we are still sticking to this announcement Aug 14 17:19:53 Right, so this article saying exactly that does not surprise me, nor does it seem shocking. Aug 14 17:19:58 no Aug 14 17:20:05 but the basic premise is flawed Aug 14 17:20:20 Is it? You guys have been upsetting developers as of late :) Aug 14 17:20:25 they compare a product available in store with a preview SDK release Aug 14 17:20:36 so yes, it's hard for developers to make money out of Android right now :) Aug 14 17:21:03 yes, that's why the article is rubbish - the comparison w/ apple is stupidly skewed Aug 14 17:21:19 tomgibara: agreed Aug 14 17:21:25 now *really* back to work :) Aug 14 17:21:50 at least we're being compared to iPhone, not Windows Mobile or Symbian Aug 14 17:22:40 hehe, i can't understand how Windows Mobile is even in the market :) Aug 14 17:22:58 as far as i can tell (having seen both) the gphone has way more power and looks much more intressting (from a developers pov). BUT: the device is too big/ugly Aug 14 17:22:59 I recently tried a WinMo touch device... Aug 14 17:22:59 it's what we buy at work, unfortunately by my recommendation, but i can't believe no one tackles them! Aug 14 17:23:11 Android isn't even likely to tackle them. With no Exchange support, everyone's hands are tied. Aug 14 17:23:15 they must have forgotten that people's fingers are slightly bigger than a stylus Aug 14 17:23:59 romainguy_ you can change to bigger keyboard, at max size of ~1cm per button Aug 14 17:24:43 alex2308: I was talking about all the widgets on the screen Aug 14 17:24:44 romainguy_: I just recently played with a Windows Mobile 6.1 touch-screen device and was baffled that by *DEFAULT* ClearType was turned off, and the font size set to "Small" Aug 14 17:24:51 scrollbars, buttons, notification icons, etc. Aug 14 17:24:51 Even I had a hard time reading the stupid thing. Aug 14 17:25:00 romainguy_ oh, yes. they definitelly suck Aug 14 17:25:05 It's like nobody takes the damn thing seriously. Aug 14 17:25:29 All I can say is I'm jealous of the Windows Mobile dev team. They must have the most cake job in the industry. Aug 14 17:26:01 The only have to make progress at a rate consistent with the software industry in 1990. Aug 14 17:26:19 the performance of their software stack is pretty scary too Aug 14 17:26:32 watching Windows Mobile rotate the screen is very entertaining Aug 14 17:26:35 Oh yeah. On this same device you could see noticable flicker as you moved around most screens Aug 14 17:26:48 well, i have to say, the idea of .NET is good, having a VM between hardware and software. so can you port your desk app to your mobile. but ... something must have went wrong ... dunno, the result sucks Aug 14 17:26:58 And yes, you can visualize how it progresses drawing a window. Aug 14 17:27:01 It does it so slow Aug 14 17:27:21 alex2308: how is that any different from J2SE/J2ME? Aug 14 17:27:44 There is still a .NET standard edition, and a Compcat Framework version for mobiles. So the same code doesn't necessarily port. Aug 14 17:27:57 its not, but afaik .net provides access to more parts of the phone, like IR Aug 14 17:28:24 So guys I'm leaving now. See you tomorrow. Aug 14 17:28:28 That's true, but the idea is not new. Their execution is just better. Android's is even better still. Aug 14 17:28:36 I haven't done any WinMo development, but based on what I hear, and what I read of the technicalities of the different platforms, I think Android has the makings of _the_ development platform of choice in the mobile arena. Aug 14 17:28:49 maybe parts wont, but my (.net 2.0) application could be ported to win mobile, i checked it. but id have to rewrite the UI Aug 14 17:29:06 true, android seems much better Aug 14 17:29:18 It's just better architected at every level Aug 14 17:29:37 but the device is big and ugly Aug 14 17:30:01 we need smaller android devices. i prefer the white TI prototype. the one that looked like a blackberry. it was so cute Aug 14 17:30:02 One development platform will spawn many phones Aug 14 17:30:07 (the devices that we dont know *notify*) Aug 14 17:30:19 You can change the phone - but not the platform so easily Aug 14 17:30:34 http://www.product-reviews.net/wp-content/userimages/2008/03/google-android-phones-fall-08.jpg Aug 14 17:30:37 this one i liked most Aug 14 17:30:39 tomgibara: that's it. Android is not some specific phone. Aug 14 17:31:06 alex2308: Android is a platform, not a phone Aug 14 17:31:22 well, there will be one phone in q4 08 Aug 14 17:32:28 Unfortunately, technical merit seems to be little correlated with commercial success, so I don't make any predictions on that. Aug 14 17:32:54 tomgibara: we're all hoping they don't fuck it up Aug 14 17:33:18 tomgibara: true, see xbox/playstation/wii Aug 14 17:33:45 alex2308: there was probably more technical merit in the Wii than in the Xbox actually :) Aug 14 17:34:36 I'm not sure that products always fail by 'fuck up' - I think there's plenty of room for chance to play a part. Aug 14 17:34:54 but the wii is (from a technical pov) very simple - and yet, its really great Aug 14 17:35:30 alex2308: that's not what I was saying Aug 14 17:35:38 I'm saying that the Xbox is a very conservative machine Aug 14 17:35:43 even though it has 3 cores Aug 14 17:36:27 AttractiveApe: you there? i will have something for you to test very soon Aug 14 17:36:29 that's why I don't think it has more technical merit than a Wii Aug 14 17:36:40 but yeah, it's more powerful when it comes to crunching numbers :)) Aug 14 17:36:42 is it? i thought its counterwise Aug 14 17:36:54 contrariwise Aug 14 17:37:15 again, Microsoft just assembled a gamer's PC Aug 14 17:37:33 whereas Nintendo had to create and mass-produce a cheap, usable, motion tracking tech Aug 14 17:37:52 in that regard the PS3 has also more technical merit than the Xbox Aug 14 17:38:06 even though games usually run at higher resolutions on the Xbox Aug 14 17:38:50 (too bad Sony fucked up on the GPU :( Aug 14 17:46:16 jasta: I'm about Aug 14 17:49:19 AttractiveApe: good :) Aug 14 18:01:53 I want an Android phone... lalala Aug 14 18:02:02 AttractiveApe: alright. i'm going to upload to android-random now. Aug 14 18:02:38 k Aug 14 18:03:23 http://phandroid.com/2008/08/14/nobody-cares-about-android/ :D Aug 14 18:04:20 AttractiveApe: ok, you need to download these two files to the device: Aug 14 18:04:30 http://android-random.googlecode.com/files/JNITest.apk to /data/app Aug 14 18:04:43 http://android-random.googlecode.com/files/libNativeMD5.so to /system/lib Aug 14 18:05:09 then fire up the JNITest app and hit Run. I need you to let me know what the 4 benchmark tests give (should be timings, x seconds elapsed). Aug 14 18:05:38 actually, wait, i hardcoded the paths of the files it reads to /sdcard, and as i recall that does not work Aug 14 18:06:06 * AttractiveApe yawns. Aug 14 18:06:24 how mcuh space do you have to work with in /data? Aug 14 18:06:47 4GB sdhc , so a bit. Aug 14 18:07:02 ahh, ok. i will need you to also create two files called /data/foo and /data/bar Aug 14 18:07:13 they should be identical, and about 4MB in size. Aug 14 18:07:26 the app will just MD5 sum them, basically. Aug 14 18:07:41 teh reason i use two files is because i don't want the kernel's read cache to interfere with benchmarks. Aug 14 18:07:48 that is also the reason why i can't write the files on the fly. they must be existing. Aug 14 18:07:56 mmmk. Aug 14 18:08:06 i created them with: Aug 14 18:08:13 jasta: "echo 3 > /proc/sys/vm/drop_caches" Aug 14 18:08:22 dd if=/dev/urandom of=foo count=4k Aug 14 18:08:29 that will drop the page-cache Aug 14 18:08:41 really? that's good to know Aug 14 18:08:45 yup. Aug 14 18:09:33 AttractiveApe: ok, i just uploaded the new JNITest apk that uses /data/foo and /data/bar Aug 14 18:09:56 so again, libNativeMD5.so must go to /system/lib, the apk in the normal place, and two identical 4MB files in /data/{foo,bar}. Aug 14 18:10:57 SanMehat: do you also know how to disable checkjni? Aug 14 18:11:07 jasta: no sorry Aug 14 18:11:10 What's the md5 on your latest JNITest.apk? Aug 14 18:11:13 i only see a way to do it from the emulator's command line, no idea how to disable it on the vogue. Aug 14 18:11:20 17112444766b439105a0567c9893b5e9 bin/JNITest.apk Aug 14 18:11:25 cool. Aug 14 18:13:51 i'll look into how to disable JNI... Aug 14 18:13:59 err, JNI checks Aug 14 18:14:02 copying the new img's to the sd card... Aug 14 18:17:59 4 tests, right jata? Aug 14 18:19:30 native foo: 0.64s, hybrid bar: 0.93s, native foo: 0.61s, hybrid bar: 0.93s Aug 14 18:19:38 yeah Aug 14 18:19:59 wow, really? Aug 14 18:20:08 I accidentally ran it twice Aug 14 18:20:11 that's the results of the 2nd run Aug 14 18:20:18 your scroll thing is broken I think Aug 14 18:20:19 did you create them with dd if=/dev/urandom? Aug 14 18:20:22 I did. Aug 14 18:20:30 no, Android's scroll thing is broken. use the keypad with it. Aug 14 18:20:36 and how big are the files? Aug 14 18:21:41 2MB each Aug 14 18:21:47 ahh ok. Aug 14 18:21:49 Want bigger? Aug 14 18:21:59 No, but I'd prefer you get it on the first run Aug 14 18:22:07 since we didn't disable the kernel's cache :) Aug 14 18:22:41 actually, try 8MB files. i'm curious if we just have a lot of unrelated overhead in a test run around half a second. Aug 14 18:23:04 then we'll be done. we've already showed that the results are at least consistent with the emulator. Aug 14 18:23:10 though it frustrates me not knowing how to turn of checkjni Aug 14 18:24:16 actually, let's just take a stab in the dark: Aug 14 18:25:13 grep for checkjni Aug 14 18:26:46 create a file called /data/local.prop Aug 14 18:26:49 and write in it: Aug 14 18:26:59 ro.kernel.android.checkjni = 1 Aug 14 18:27:14 and also modify default.txt and add android.checkjni=1 to the cmdline Aug 14 18:27:23 (who knows, one of them might take *grin*) Aug 14 18:28:11 for this run, native foo: 4.61s, hybrid bar: 4.29s, native foo: 2.35s, hybrid bar: 3.64s Aug 14 18:28:31 8MB foo / bar size from urandom Aug 14 18:28:52 that first test doesn't seem quite right ;) Aug 14 18:29:41 the native method is guaranteed to take less cpu time hehe. Aug 14 18:29:44 are the first two hashing and the second two vanilla? Aug 14 18:30:00 both tests are the exact same thing, i just run them twice to attempt to crudely detect anomalies Aug 14 18:30:12 and also observe the effect of the page cache if it's being used. Aug 14 18:30:32 in this case, that second test was more what i'd expect. Aug 14 18:30:42 but still, we won't know for sure until checkjni is off. Aug 14 18:30:56 perhaps the device was busy on the first test Aug 14 18:30:56 AttractiveApe: try one more time modifying those two files for checkjni and run again. we'll see if we can get different results. Aug 14 18:31:21 tomgibara: what's interested here is that the device's performance is dominating the emulator. Aug 14 18:31:43 It's not a bad piece of hardware :p Aug 14 18:31:58 i get about 2.4s and 3.5s running on the emulator for 4MB files. Aug 14 18:33:04 AttractiveApe: it's an MSM7500 400MHz. i hear rumors that the first android device is expected to be in the 500MHz range ;) Aug 14 18:33:16 jasta: probably 520Mhz Aug 14 18:33:25 <_Marvin_> hey Aug 14 18:33:30 jasta: I'm not that surprised Aug 14 18:33:39 jasta: the new touch pro / diamond's have those procs. Aug 14 18:34:14 apparently the diamond's were just released for buying today. Aug 14 18:35:35 tomgibara: i would not have expected the device to win out in number crunching and disk I/O. Aug 14 18:37:18 I would expect it to dominate in anything reads/writes a lot to memory Aug 14 18:37:29 since there were comments from a Google engineer about the slowness of the memory emulation (iirc) Aug 14 18:38:18 AttractiveApe: if you get time, try it once more to see if we can get checkjni off Aug 14 18:38:28 if not, we'll just leave it here. the results i got were otherwise consistent with the emulator. Aug 14 18:39:17 Though I'm not well informed enough to form a worthwhile opinion about the performance. Aug 14 18:41:57 AttractiveApe, where can you buy a diamond? Aug 14 18:42:10 uh, HTC diamond Aug 14 18:42:15 Telus Aug 14 18:42:21 They just opened it today. Aug 14 18:42:31 is that Europe or Asia? Aug 14 18:42:31 jasta: surely that checkjni is going to skew the results though Aug 14 18:42:38 tethridge: Canada, so North America. Aug 14 18:42:46 tethridge: the CDMA version, not GSM. Aug 14 18:43:05 I'd love to Sprint have something like that for Android Aug 14 18:43:12 it could happen. :-) Aug 14 18:43:26 <_Marvin_> I'd like it too Aug 14 18:43:45 it would make it nice with their $15 a month unlimited data plane Aug 14 18:43:47 plan Aug 14 18:45:23 argh, seems like my app is getting to big, as the debugger likes to get fucked up when I go behind my menu o_O Aug 14 18:46:27 <_Marvin_> how big is it? Aug 14 18:46:54 just 25k lines but a lot on a MapActivity Aug 14 18:57:15 i hope you dont have one of those 10k line files :) Aug 14 18:57:33 nono ^^ Aug 14 18:57:50 Overlay 1k (needs 40ms to be drawn) Aug 14 18:57:56 and some simple calculation. Aug 14 18:58:04 works really smooth when "Run"ning Aug 14 19:48:39 plusminus_: have you tried it on a real phone yet? Aug 14 20:02:02 jasta: I have non Aug 14 20:02:14 just the ZoomMDK Aug 14 20:02:41 and the Android-Support is in Korean Aug 14 20:02:45 :D Aug 14 21:27:17 romainguy_: is Android really totally resolution-independent? Aug 14 21:27:48 as in, arbitrary resolutions will implicitly be supported? Aug 14 21:28:21 that would be very excited to homebrew projects like custom in-dash car stereoes or touch-screen kiosks. Aug 14 21:33:27 resolution-independent is more than that Aug 14 21:33:31 but yes, Android doesn't care Aug 14 21:33:36 applications though must be careful Aug 14 21:34:03 of course, we've already run into that numerous times with the Vogue :) Aug 14 21:36:19 romainguy_: is the home screen currently replaceable, or is that a 1.0 thing? Aug 14 21:36:31 by current I mean M5 Aug 14 21:36:57 it's replaceable Aug 14 21:37:03 the next SDK will ship with a sample Home Aug 14 21:37:12 really? that's cool. Aug 14 21:37:24 actually the sample Home will be the one you have in M5 Aug 14 21:37:28 although if you replaced Home you would break my heart :) Aug 14 21:37:33 i was actually going to comment that the SDK should do a better job demonstrating more advanced functionality Aug 14 21:37:42 well, your Home screen sucks bad on the Vogue :) Aug 14 21:37:47 the M5 one that is Aug 14 21:37:54 i know the 85772 is different *wink* Aug 14 21:38:11 ah yes, the M5 one totally sucks Aug 14 21:38:34 on the Vogue, the M5 one reverts to showing only a single column expanded by the "All" button Aug 14 21:38:38 yep Aug 14 21:38:48 at least it still works :P Aug 14 21:38:49 so you have one long scrolling column on the left with all apps, and one long column on the right with recent/defaults Aug 14 21:38:54 jasta: i know the 85772 is different *wink* << have you tried it? Aug 14 21:39:00 yeah, i was actually surprised it was so flexible to even work. Aug 14 21:39:23 M5 Home is interesting as a sample Aug 14 21:39:26 romainguy_: maybe, maybe not ;) Aug 14 21:39:43 jasta: do you always have to be a punk-ass? Aug 14 21:39:46 it shows how to use custom layout, custom attributes, layout animations, animations, etc. Aug 14 21:39:46 :) Aug 14 21:40:12 using Android on the Vogue has made it a lot easier to understand how to create layouts that are flexible. Aug 14 21:40:35 usually supporting landscape and portrait is good enough to support multiple resolution Aug 14 21:40:57 and apps can always use layout-320x240 or whatever they see fit Aug 14 21:41:13 romainguy_: what i like is that it is easy to expand from there. what i mean is that you can initially create layouts that work in portrait and landscape, then you can fine-tune as your app matures. Aug 14 21:41:43 for example, my solitaire game currently works on any resolution, but certain resolutions should use a custom layout to better utilize available space. Aug 14 21:44:02 one thing i don't like though is that because of all this magic in the directory structure you can't unclutter your layouts by creating your own hierarchy Aug 14 21:44:10 so large apps have tons of xml layout files in the same dir. Aug 14 21:44:43 after 1.0 Aug 14 21:48:57 romainguy_: What do the layout xml files get processed into, bytecode? Aug 14 21:49:06 no, binary XML Aug 14 21:49:13 tomgibara: they go into a binary XML format, which i have loosely documented. Aug 14 21:49:45 jasta: Oh yeah, I forgot about that Aug 14 21:49:47 tomgibara: i wrote a tool ages ago that reverted them back to plain XML Aug 14 21:50:33 bytecode would seem a good choice though Aug 14 21:51:03 Not really. Aug 14 21:51:06 not compact enough for this purpose perhaps Aug 14 21:51:17 why not really? Aug 14 21:52:09 Well think about it, you could either explicitly generate instructions (as in a bytecode) to pull parse, then create an interpretor for those instructions that runs atop another interpreter (Dalvik). Aug 14 21:52:42 Or you could simply construct code that can parse directly from the file stream, which javac will generate to a bytecode. You just removed one layer of indirection. Aug 14 21:52:54 No I meant Dalvik bytecodes, basically each layout file would become a function that operated on some context Aug 14 21:52:55 At the cost of some flexibility, which you won't need. Aug 14 21:53:32 Honestly, that would seem like excessive optimization, with dubious benefits. Aug 14 21:54:16 well compactness of representation would be my main concern - but I certainly wouldn't regard it as excessive optimization Aug 14 21:54:17 it would also be annoying to maintain Aug 14 21:54:40 tomgibara: actually, the main design goal was parser performance, not file size. Aug 14 21:54:58 I don't see that it's any more difficult to maintain than a parser Aug 14 21:55:07 everything is word-aligned, for example, so that they can just map most of it. Aug 14 21:55:11 well it produces XML that are tied to the API Aug 14 21:55:21 since you generated bytecode Aug 14 21:55:45 Yes, I see that, but the parser is tied to the API too. Aug 14 21:55:58 Or do you mean for compatibility of apks between sdks Aug 14 21:56:02 except that the compiled files are not Aug 14 21:56:27 tomgibara: by the way, the XML parser Android has is not being run by Dalvik, so your approach would actually remove an opt possibility. Aug 14 21:56:31 besides, we know have include and merges and other funky stuff that are much easier to handle in the parser Aug 14 21:56:37 it was written in C++, and is exposed through JNI. Aug 14 21:56:49 and given the size of the XML documents we have, I doubt that compiling to bytecode would be a big improvement Aug 14 21:57:02 most of the time spent creating layouts is spent executing View's constructor Aug 14 21:57:06 the parsing part is nothing Aug 14 21:57:14 fair enough, I naturally defer to your better understanding Aug 14 21:57:33 tomgibara: do note my comment above ;) Aug 14 21:57:58 you would still be adding a layer of indirection even by parsing directly to Dalvik, as it is not even written in Java for the most part ;) Aug 14 21:58:08 jasta: I don't see that as particularly valid Aug 14 21:58:38 A 'dexified' version of the layout would basically be calling a bunch of constructors and add methods Aug 14 21:59:24 I can't imagine any significant improvement by moving that work into C Aug 14 22:00:13 tomgibara: then you're not thinking clearly. for example, in C you can map the internal file formats data structures directly to structures in C, without constructing anything other than the buffer you needed to read it. Aug 14 22:00:23 that's why the file format is aligned. Aug 14 22:00:35 That the parsing is very little overhead, is the key reason - and I note again following on from a previous discussion - not written in Java Aug 14 22:00:38 it's non-portable, of course, but it is a common optimization technique. Aug 14 22:00:54 I'm familiar with that technique Aug 14 22:01:08 but it doesn't work when building a bunch of Java objects Aug 14 22:02:00 that is true, it would still have to make a copy to get it into Java. Aug 14 22:02:40 it would have to do more - allocating objects on the heap invoking the constructors etc. Aug 14 22:02:50 However your approach would not be any different than shaving off a few CPU instructions in this parser loop, doing nothing to tear down all the work doing allocations and copies. Aug 14 22:03:07 exactly, that's why its a bad idead Aug 14 22:03:09 *idea Aug 14 22:03:25 because the parsing isn't the main overhead Aug 14 22:03:55 I'm quite happy to have bad ideas! As long as I get to understand why they're bad :) Aug 14 22:04:04 however, i just recently demonstrated that letting Java handle file I/O is some margin slower on Dalvik ;) Aug 14 22:04:18 so right there you score a win with the native C implementation hehe Aug 14 22:04:20 which is why they're using C to do the parsing Aug 14 22:04:32 but with a dexified version there is no parsing Aug 14 22:04:49 it's the closest equivalent to memory mapping the c structures you were talking about Aug 14 22:04:50 now, i remain skeptical about my own results. That Sun's JVM shows no significant difference but Dalvik's does has me doubting that Dalvik is going as far as it can to opt these calls. Aug 14 22:05:02 also, i can't seem to turn checkjni off. Aug 14 22:05:05 unless checkjni is interfering Aug 14 22:05:21 well, i initially thought i had it off by passing -nojni to the emulator Aug 14 22:05:42 but upon closer inspection, i don't think the flag works as documented. i can't find any proof demonstrating that it was off, and the benchmark results are unaffected. Aug 14 22:06:03 You need to pass a flag to the dalvik executable Aug 14 22:06:06 so either it is insignificant (unlikely) or something isn't working as i expect. Aug 14 22:06:14 or put some magic property into some file somewhere Aug 14 22:06:42 yeah, it needs closer inspection Aug 14 22:06:59 -Xcheck:jni is command line switch to Dalvik Aug 14 22:07:02 what's frustrating is that it doesn't tell you anywher e:) Aug 14 22:07:18 tomgibara: well, read emulator -h (-nojni) Aug 14 22:07:25 i thought that was it, but as i said, i'm not so sure. Aug 14 22:08:13 is it possible to grep the device files for checkjni or -Xcheck:jni or something like that? Aug 14 22:08:18 i did Aug 14 22:08:36 i took a look at the kernel's command line params with and without -nojni Aug 14 22:08:41 i also looked at adb logcat in both cases Aug 14 22:08:54 grep -i jni, even. didn't find anything useful. Aug 14 22:09:03 hmm :( Aug 14 22:09:35 i found some site somewhere that suggested ro.kernel.android.checkjni=0 would turn it off, and i think that file is either initrd:/etc/default.prop or /data/user.prop or something Aug 14 22:09:57 /data/local.prop Aug 14 22:10:12 I love switch names like -nojni which actually mean make jni faster :) Aug 14 22:10:18 i had been meaning to confirm that suspicion, but i haven't yet. Aug 14 22:10:30 tomgibara: yeah, the name of the switch is largely why i think it doesn't work. Aug 14 22:10:36 because some goober intern seems to have created it. Aug 14 22:10:43 hehe Aug 14 22:10:56 You don't have much faith in people :) Aug 14 22:11:11 that there's no log line saying "JNI checking is off!" makes me too lazy to test my above suspicions ;) Aug 14 22:11:17 since i'd have to use performance benchmarks to test, and that sucks Aug 14 22:11:45 tomgibara: well, or i don't have much faith in people writing software. and i don't. humans are not good at this vocation :) Aug 14 22:12:58 * ianedgerly greets all Aug 14 22:13:02 hello Aug 14 22:13:29 jasta: sadly, I believe that to be very true Aug 14 22:14:04 tomgibara: it's simply against our grain. it's not how we were meant to think. Aug 14 22:14:37 yeah Aug 14 22:16:22 jasta: I think partially, it's because we don't have the tools yet. Aug 14 22:17:02 Mathematics is possibly the most 'unhuman' subject Aug 14 22:17:24 but the abstractions that have accrued over several millenia make it more approachable Aug 14 22:17:41 Einstein has this great quote that i really love, hang on let me find it Aug 14 22:17:50 it popped into my head when you said that hehe Aug 14 22:18:05 about his problems with mathematics? Aug 14 22:18:16 no, that is a tired quote Aug 14 22:18:34 I'm glad you said that :-) Aug 14 22:19:15 Einstein has lots of gems but that wasn't one of them. Aug 14 22:20:38 For example (not the one I was looking for): "Only two things are infinite, the univers and human stupidity, and I'm not sure about the former." Aug 14 22:20:56 but hang on, i'll find the one i'm thinking of. i'm determined now :) Aug 14 22:24:40 Here it is: "As far as the laws of mathematics refer to reality, they are not certain, as far as they are certain, they do not refer to reality." Aug 14 22:25:13 Yes, very accurate phrasing there. Aug 14 22:25:15 there's another one i was thinking of too but now i'm bored with this search and won't find it. Aug 14 22:26:27 Einstein was really an interesting fellow. I like reading up on his life and such. Aug 14 22:26:49 A passive curiousity of mine is to explore the personalities of some of the "great" well known scientists throughout history. Aug 14 22:29:12 One of the things I find amusing about einstein is that it seems to be a persistent myth that he was bad at mathematics Aug 14 22:29:21 Gallileo's story tickles me. I like to reiterate it to the religiously oppressive folks in my life. Aug 14 22:29:59 I'm sure there are a lot of myths about Einstein. Most folks have absolutely no idea who he was. Aug 14 22:30:02 I doubt if it tickled Gallileo very much Aug 14 22:30:02 None whatsoever. Aug 14 22:30:36 good evening everyone. Aug 14 22:30:47 Well, I suspect most have heard of him Aug 14 22:30:48 tomgibara: I just meant the part about being put under house arrest because they felt bad killing an old man, and then he went on to write his best work published by a friend. Aug 14 22:30:56 Oh yes Aug 14 22:31:02 tomgibara: yeah, heard of him, but they have literally no clue what he did. Aug 14 22:31:31 Maybe Gallileo's still laughing at them now from 'the other place' Aug 14 22:33:09 jasta: It's a pity because Einstein's thought experiments are some of the most accessible elements of modern science Aug 14 22:34:07 Though there's a sad irony that he received his nobel prize for a key discovery in quantum theory which he spent so much time trying to disprove Aug 14 22:34:36 That's how science work. There's nothing sad about that to me. Aug 14 22:35:08 Some of the greatest personal epiphanies I've had came from "finally" bothering to prove some long-held believe to find I was quite wrong all along. Aug 14 22:35:15 belief* Aug 14 22:35:36 What is sad is that he refused to believe it was true, even after helping to prove it Aug 14 22:36:53 i have wasted half my afternoon when i should have been coding. Aug 14 22:36:57 hehe Aug 14 22:37:01 i'm going to get back to it :) Aug 14 22:37:13 okay happy coding Aug 14 22:42:02 tomgibara: Einstein was "bad" at mathematics because he went to school in Switzerland, where "6" is best, right? Aug 14 22:45:01 Are you testing my german :-) Aug 14 23:09:54 do I? Aug 14 23:10:42 ? perhaps I didn't understand your message Aug 14 23:13:55 You were talkign about Einsteins bad math grades and I wrote what I had in my mind... Aug 14 23:15:30 So 6 is the best grade? Aug 14 23:15:45 or 5, dunno (in Switzerland) Aug 14 23:15:52 in Germany its the other way around Aug 14 23:15:59 (1 is best) Aug 14 23:16:07 back to android now Aug 14 23:16:15 MoseyCode will be free, I guess? Aug 14 23:16:35 but did you think about preotecting an app from privacy? Aug 14 23:17:03 as when you want to sell it you need some key-code (or IMEI based) security... Aug 14 23:17:29 sup +/- Aug 14 23:17:38 ? Aug 14 23:18:20 I hope something like that will come with the next sdk or with T-Mobile's Android-Market... Aug 14 23:19:17 IMEI will probably be a reliable way of uniquely identifying users, but you need to be careful not to expose it Aug 14 23:19:34 yeah... Aug 14 23:19:47 Moseycode needs accounts so that published barcodes are 'accounted for' Aug 14 23:20:03 imo its so much overhead, if every developer has to create a solution for that on its own :( Aug 14 23:20:06 but it lodges a securely generated key on the device and uses that Aug 14 23:21:04 my plan is to, in time, adopt openID, that will help some users to avoid being overloaded with accounts Aug 14 23:21:32 There's no reason why you can't use, the google authentication APIs Aug 14 23:21:58 It will of course require that users need to sign up to a google account to use your app Aug 14 23:22:09 but if the rumours are true :)... Aug 14 23:46:17 I was thinking of something like this: Aug 14 23:46:52 onCreate(Bundle b){ Aug 14 23:46:54 .... Aug 14 23:46:59 boolean verified = this.verifyThisAppToAndroidMarketByIMEI(); Aug 14 23:47:04 if(!verified) Aug 14 23:47:09 this.finish(); Aug 14 23:47:13 else Aug 14 23:47:18 this.setContentView(...); Aug 14 23:47:19 ... Aug 14 23:47:20 } Aug 14 23:50:41 plusminus_: ...contact the network each time your application opens? no, no. :) Aug 14 23:51:13 that was the first thing that came to my mind @ 2am :/ Aug 14 23:51:29 hahhaah Aug 14 23:51:32 but I think you got what I mean Aug 14 23:51:33 in general, you must always regard the network as slow, unstable and possibly just totally gone. Aug 14 23:51:54 I would say totally gone, unstable and then slow :)) Aug 14 23:51:55 yeah just throw it in a runnable in a new thread. if it works it works, if not, that's life :P Aug 14 23:52:01 i'm just talking on the basic idea Aug 14 23:52:39 or how do you manage Anti-Piracy ? Aug 14 23:52:51 plusminus_: still, if something like that was possible (god i hope they don't do that to us), then it would be done with a periodic phone-home check. Aug 14 23:52:56 nothing that runs each time your app runs. Aug 14 23:53:16 yeah probably. Aug 14 23:53:19 plusminus_: Most mobile software does not even attempt to thwart piracy. Doing so would damage paying customers user experience for the most part. Aug 14 23:53:53 What I fear is that I have to send out keys or something o_O Aug 14 23:54:13 when someone buys my app in the "Market" i won't recognize that at all Aug 14 23:54:16 plusminus_: most likely work with operator on that, or build security enforcement into your backend Aug 14 23:55:45 Imo the "Market"-company should assure that only the user who bought my app is allowed to use it. Aug 14 23:56:23 thats my opinion Aug 14 23:56:27 well, it needs to be technically feasible too Aug 14 23:56:30 the enforcement of it Aug 14 23:56:35 but dont spend too much effort Aug 14 23:56:38 mobile piracy is still low Aug 14 23:56:43 not worth the effort imho Aug 14 23:56:54 just put some easy stopgap measures to keep out the riff raff Aug 14 23:57:13 experienced crackers will get through it no matter what Aug 14 23:57:27 yeah, I just don't want it to get copy-pastes Aug 14 23:59:19 i'm sure someone will write a disassembler for dex Aug 14 23:59:23 and then trivially crack programs :) Aug 14 23:59:43 just like they do with PC software. Aug 15 00:00:06 (for any of you that have never done it, it's very simple.) Aug 15 00:00:15 That selling/anti-piracy-thing is really new for my, I'm 100% unexperienced... that sux. Aug 15 00:00:39 anti-piracy is largely a waste of time. price your product correctly and you won't see a problem. Aug 15 00:00:51 price your product absurdly high and you will see it happen rampantly, and you won't be able to stop it. Aug 15 00:02:21 hm 10$ are worth looking for a crack? Aug 15 00:02:38 if 100 people do so, thats hell of money for me!! Aug 15 00:02:42 don't ask me. Aug 15 00:08:48 piracy is not necessarily a negative - look at the rise of Microsoft Aug 15 00:09:20 ^^ you mean like, it spreads like a virus? Aug 15 00:10:30 I mean that it can entrench a product and make it a defacto standard Aug 15 00:10:46 I also think there's an odd psychology present Aug 15 00:11:00 hm good point Aug 15 00:11:16 People will value an illegal copy of a piece of expensive software more than they will a legal copy of free software Aug 15 00:11:26 even if the free software is better in every regard Aug 15 00:11:44 bc of being the bad guy Aug 15 00:11:46 ^^ Aug 15 00:11:49 somehow they feel as though they have 'gained' the monetary value of the copied software Aug 15 00:12:02 even though they have done no such thing of course Aug 15 00:12:27 the thing i fear is of that is: Aug 15 00:12:41 Someone got my app hacked/cracked whatever Aug 15 00:12:50 which will happen Aug 15 00:13:13 when his friends have it, will he give them the cracked version, will they buy it then? Aug 15 00:13:20 the probably never buy it Aug 15 00:13:23 no they won't Aug 15 00:13:42 I think it comes down to a cost benefit analysis Aug 15 00:13:45 add some exponential growth to it Aug 15 00:13:51 probably Aug 15 00:13:55 As jasta was hinting at: Aug 15 00:14:27 Everything you add to make the sofwtare harder to copy will almost certainly have a negative impact on your paying users Aug 15 00:14:48 if the paying user does not see it= Aug 15 00:14:49 ? Aug 15 00:15:06 There is a trade off there - and in my opinion it should be set very very much in the paying customers favour Aug 15 00:15:16 What don't they see? Aug 15 00:15:30 Extra network chatter, extra battery usage, extra code Aug 15 00:15:44 everything you chose to do with an application has an impact Aug 15 00:16:20 not to mention things like registration/logins etc. which have a direct impact on the users Aug 15 00:16:22 even in terms of your time: the time you spend on these features is time you don't spend on better things Aug 15 00:16:28 that oo Aug 15 00:16:30 *too Aug 15 00:16:46 also, again, do realize how easy piracy is even with the industry's leading efforts. Aug 15 00:17:08 Again, in my case, I needed to make a registration system for Moseycode because ultimately the codes are a finite resource Aug 15 00:17:15 But even then I agonized over it Aug 15 00:17:41 hm ok, so you would guess something simple that hurts the user at a minimum (bets would be nto at all) Aug 15 00:17:52 Anything that gets between the user and the task they want to achieve with your application is a big negative Aug 15 00:18:48 Well, this strays slightly into analyzing the "business model" Aug 15 00:19:27 As soon as you are looking to raise revenue from an application, you are looking to take something away from the users of that application Aug 15 00:20:00 and you need to form a proposition that is paletable to the users Aug 15 00:20:47 This ignores the complication that you need to analyze the market and segment it Aug 15 00:21:09 thats what I don't have skills/time for Aug 15 00:21:26 It's good that you recognize that Aug 15 00:21:44 So the answer is probably not to sweat it Aug 15 00:21:56 But don't take that as cast iron business advice! Aug 15 00:22:04 Do you have employees? Aug 15 00:22:15 Do you have shareholders? Aug 15 00:22:15 ^^ only my cat Aug 15 00:22:21 Do you need the money to live? Aug 15 00:22:26 nope Aug 15 00:22:43 Then absolutely don't stress about it :) Aug 15 00:22:52 but it could be a lot and i don't want to worry about it later on Aug 15 00:23:25 My experience is that getting users - any users, let alone paying users - can be quite a struggle Aug 15 00:23:35 so be grateful when someone pirates your application Aug 15 00:23:41 :D Aug 15 00:23:41 it's free marketting Aug 15 00:24:35 It increases the number of people who are interested in seeing it improve (which won't be all the people who copy it, most probably won't even use it) Aug 15 00:25:03 It means that when you bring out the next version, more people might be interested in upgrading Aug 15 00:25:33 hm so maybe think of anti-piracy later on. Aug 15 00:25:57 There's a real irony about piracy Aug 15 00:26:45 Let it be 100 Million Android-Phones. I think 10% would use it for navigation. and 1% of them chooses my app. Thats still 100000 customers. Aug 15 00:27:03 *stunning* Aug 15 00:27:07 Suppose your application is wildly popular and copied a great deal - you might find yourself owning the most popular mobile phone sat-nav application in Germany Aug 15 00:27:12 you think thats realistic ? Aug 15 00:27:18 That makes it extremely valuable Aug 15 00:28:17 The piracy (within reason) ends up making many application more valuable Aug 15 00:28:51 realistic: not on a short time scale Aug 15 00:30:07 Again, please take this only as an informal suggestion, but you should sell the software through a limited company (or the german equivalent) Aug 15 00:30:13 Gmbh? Aug 15 00:30:54 think so Aug 15 00:31:01 The first person who crashes into a bollard because your application crashed might sue Aug 15 00:31:08 I'd need 25000 € if I remember right Aug 15 00:31:26 wow, it's a few hundred pounds in the UK Aug 15 00:32:23 "Under German law, the GmbH must have a minimum founding capital of €25,000, of which 25% or at least €12,500 has to be contributed by its members..." Aug 15 00:33:22 Got a idea from the groups: Aug 15 00:33:29 "Another approach would be if some Android app stores signed the APKs based on the IMEI, so if you buy it from the store, you can only use your downloaded APK on that device." Aug 15 00:33:52 thats the kind of complexity I've been looking for :) Aug 15 00:34:18 easy and probably still crackable Aug 15 00:34:22 yes as jasta pointed out, that can easily be cracked Aug 15 00:35:01 and you're going to really annoy the users who's phones breakdown and need to be repaired Aug 15 00:35:27 there will be solutions for that Aug 15 00:35:47 as the 'Store' will remember that you bought the app Aug 15 00:35:51 and probably dozens of other situations could arise Aug 15 00:36:05 which means I need an account at the store? Aug 15 00:36:45 you surely need to get charged to buy any app, what means you need an account Aug 15 00:36:52 .. at the store Aug 15 00:37:13 Not if you used paypal (just as an example, I'm not advocating it) Aug 15 00:37:34 I don't think there will be a anonymous way (whereas it is possible of cours) Aug 15 00:38:04 I don't think T-Mobile will leave out that chance... Aug 15 00:38:08 yes, so now I need a username (or my mobile no.?), a password, an individually 'signed' apk - tied to one phone... Aug 15 00:38:15 (whatever the other carries will do) Aug 15 00:38:58 :( I just want rais anti-piracy above a copy-paste complexity :( Aug 15 00:39:27 :) Every comercial software vendor does Aug 15 00:39:52 This is a key reason why the iPhone app store has been such a success Aug 15 00:40:19 It's simple for users and protective of the application distribution Aug 15 00:41:10 what about installing a Appstore-App on a "repaired phone" Aug 15 00:41:52 I assume that the user has an account with apple via the activation process Aug 15 00:42:07 the new phone gets activated by the user Aug 15 00:42:20 the purchased applications are tied to the account Aug 15 00:43:23 The point is that the account is a single aggrevation for user that then aggregates all the apps Aug 15 00:43:38 thats pretty much the same as the signed-apk idea, which is also bound to the store-account Aug 15 00:43:55 hm maybe we will have something similar too Aug 15 00:43:57 ? Aug 15 00:44:52 Yes, perhaps, I'm happy to just wait and see what's unveiled Aug 15 00:45:22 I've got to get some sleep now ^^ 3am is approaching Aug 15 00:45:32 thx for the thoughts :) Aug 15 00:45:34 me too, gn Aug 15 00:45:37 np Aug 15 00:45:40 gn Aug 15 02:19:59 sigh my motorola q is toast it looks like :( Aug 15 02:48:55 working on rounding out my solitaire game so that it's actually a product. Aug 15 02:49:11 hopefully Google didn't also write a better solitaire engine :) **** ENDING LOGGING AT Fri Aug 15 02:59:56 2008