[00:44:18] <mizzle>	 how to open the file menu on the new crunchbang ??
[10:30:42] <yurik>	 anybody alive and well here?
[10:30:58] <yurik>	 have a silly question about the average page size
[10:31:20] <yurik>	 e.g. what's the average (compressed) download size of a page?
[10:31:37] <yurik>	 including images etc
[11:49:14] <yurik>	 anybody?
[12:25:48] <qchris>	 yurik: I ran a brief (naive) test on today's sampled-1000 logs, and there the average seems to be 8875 bytes per request
[12:26:27] <qchris>	 but that includes redirects, and all kinds of strange requests.
[12:27:10] <qchris>	 Also, for some responses header's are included in the count, for some they are not.
[12:27:16] <yurik>	 qchris, yes, i was running the same stuff on those logs, but the problem is that it does not include any image/js stuff
[12:27:56] <yurik>	 i was only considering 200 status
[12:28:33] <qchris>	 Mhmm. I am not aware of a collection that contains more data than sampled-1000.
[12:28:36] <yurik>	 resulting in 10-27 KB per page
[12:29:05] <yurik>	 qchris, i suspect that it does contain it, but varnish is not marking them with the zero= tag
[12:29:22] <yurik>	 all the images / javascripts might not be accounted for
[12:30:42] <qchris>	 Zero tagging is happening only on the text varnishes isn't it?
[12:30:54] <qchris>	 Or rather the mobile ones.
[12:35:28] <qchris>	 yurik: I just checked, it's just the mobile caches.
[12:36:25] <qchris>	 But isn't that inherent by the way it is set up?
[12:36:27] <yurik>	 qchris, yes, but i need to give some stats to our biz dev people - they need them to talk to partners
[12:36:47] <qchris>	 ok.
[12:36:53] <yurik>	 its a sales thing - we need to explain to the partner how much data we are talking about
[12:37:19] <yurik>	 so that they can say yei/nei for zero :)
[12:37:24] <qchris>	 :-)
[12:37:26] <yurik>	 (hopefully yei)
[12:37:43] <qchris>	 If there is some concrete measurement you need to get the to say yei, let me know.
[12:37:53] <qchris>	 s/the to/them to/
[12:40:10] <yurik>	 qchris, well, the total page weight is the measurement they need
[12:40:21] <yurik>	 which means pagesize + js + css + media
[12:40:44] <yurik>	 images would be the hardest to average
[12:41:05] <qchris>	 broken down by pagesize, js, css, media or just the total sum?
[12:41:07] <yurik>	 and all this is preferably per zero carrier, but for now we can use global average i guess
[12:41:12] <yurik>	 total is fine
[12:41:17] <yurik>	 per carrier :)
[12:41:25] <yurik>	 if possible
[12:41:30] <yurik>	 if not - average overall
[12:41:40] <qchris>	 We do not have that data per carrier.
[12:41:48] <qchris>	 Not even by "zero" / "non-zero"
[12:41:57] <yurik>	 do we know if the request came from mobile/non-mobile?
[12:42:30] <qchris>	 if by mobile you mean "mobile site" instead of "mobile device", then we know that for some portions of the traffic. But not in general.
[12:43:22] <yurik>	 do we know average media (jpeg/png/...) size as requested from mobile devices
[12:43:28] <qchris>	 No.
[12:43:32] <yurik>	 bummer
[12:43:35] <qchris>	 At least I do not know it.
[12:43:58] <qchris>	 Average size has not been of interest to us up to now as far as I know.
[12:44:11] <qchris>	 By when do you need this data?
[12:44:18] <yurik>	 yesterday :)
[12:44:21] <qchris>	 :-D
[12:44:27] <qchris>	 Ok. Hand over the time machine.
[12:44:40] <yurik>	 at least biz devs have been asking for it for a while since that is the biggest sticking point of signing up to zero
[12:44:55] <yurik>	 carriers want to know how much data they will donate
[12:45:10] <yurik>	 and choose images/no-images option
[12:45:19] <qchris>	 So I assume they also need to know by zerodot vs mdot?
[12:45:22] <qchris>	 Yes.
[12:45:25] <yurik>	 not really
[12:45:32] <yurik>	 because we are obsoleeting zero
[12:45:42] <yurik>	 all new carriers will (hopefully) be m. only
[12:45:47] <yurik>	 via IP
[12:45:50] <qchris>	 Ok.
[12:46:04] <yurik>	 so no need to even specify that there is zero.
[12:46:09] <qchris>	 Ok.
[12:46:52] <qchris>	 And it's mostly used for "back of the envelope" computations. Right?
[12:49:05] <qchris>	 I'll try to come up with some numbers. But that will take some time.
[12:49:34] <qchris>	 I assume that's more pressing than getting the remaining dashboards up?
[12:49:36] <yurik>	 qchris, thank you!!! I already have some breakdown for the main html portion, so only images are really needed.
[12:49:44] <yurik>	 hmm.. no idea :)
[12:50:04] <qchris>	 If it's not more pressing than that, then I'll have to push back.
[12:50:15] <qchris>	 Because that was the part I was about to start today :-(
[12:50:47] <yurik>	 qchris, i will ask dan today in 5 hours
[12:50:51] <qchris>	 Ok.
[12:50:55] <yurik>	 he should be the one to set priorities
[12:51:01] <qchris>	 Ok.
[12:51:11] <yurik>	 i have downloaded all the sampled html files for jan & feb to my machine
[12:51:24] <yurik>	 will try to get all the image tags out and evaluate their sizes
[12:51:35] <yurik>	 if wiki crashes, it wasn't me
[12:52:01] <yurik>	 that should give me some ballpark figures already, so should be ok
[12:53:53] <qchris>	 Ok. One request less on our shoulders. Even better :-D