| 00:14.40 | *** join/#uphpu synic (n=synic@c-67-177-16-249.hsd1.ut.comcast.net) |
| 00:48.55 | *** join/#uphpu wps (n=wade@65.73.113.199) |
| 00:48.59 | *** mode/#uphpu [+o wps] by ChanServ |
| 00:55.44 | *** join/#uphpu tierra (n=tierra@c-67-182-244-39.hsd1.ut.comcast.net) |
| 01:00.25 | *** join/#uphpu synic (n=synic@c-67-177-16-249.hsd1.ut.comcast.net) |
| 04:59.38 | *** join/#uphpu bigdog_ut (n=bigdog_u@shylonhunter.dsl.xmission.com) |
| 04:59.38 | *** mode/#uphpu [+o bigdog_ut] by ChanServ |
| 05:00.45 | bigdog_ut | whatz up all |
| 05:31.54 | diskotec|w | http://www.harveydanger.com/ <--protesting RIAA filesharing crackdown by releasing their latest album on their website...free for the download. they sound ok I guess. I like the first song. |
| 05:37.06 | diskotec|w | bigdog: I posted this earlier then I noticed you weren't here...free gentoo support for the next two weeks (thing is they could probably use your help...not that you'd help for free though)--> http://www.gen-ux.com/node/16 |
| 05:37.33 | diskotec|w | well good night |
| 13:59.52 | mindjuju | ~philosophize |
| 13:59.53 | jbot | hmm... philosophize is this... i'd rather have a bottle in front of me then a frontal labotomy |
| 13:59.53 | *** join/#uphpu wps (n=wade@66.237.80.85) |
| 13:59.55 | *** mode/#uphpu [+o wps] by ChanServ |
| 14:31.22 | *** join/#uphpu synic (n=synic@66.239.17.226.ptr.us.xo.net) |
| 14:59.09 | mindjuju | anybody around? |
| 14:59.46 | RyanE | nope. |
| 14:59.53 | mindjuju | :) |
| 14:59.58 | mindjuju | k, here is a quick question |
| 15:00.18 | mindjuju | let's say you have a form that allows clients to update their membership records |
| 15:01.10 | mindjuju | in the server script, do you poll all the form fields and update the sql table regardless of change or is it more efficent to try to determine which fields were changed and only update sql fields that have been changed |
| 15:01.30 | wps | it'd change them all |
| 15:01.50 | mindjuju | yeah, that's what i've been doing, but then started to wonder about efficency |
| 15:02.07 | mindjuju | towards changing them all, there is a lot of overhead required to determine if the field changed or not |
| 15:02.18 | dataw0lf | little or no benefit |
| 15:02.18 | mindjuju | seems like it'd be faster, to just update everything |
| 15:02.21 | dataw0lf | might even be worse |
| 15:02.30 | wps | yeah, just write over everything |
| 15:02.34 | dataw0lf | (forcing yourself to check whether they've been updated or not) |
| 15:02.53 | dataw0lf | but it's really dependent on specific variables |
| 15:02.58 | dataw0lf | your SQL structure |
| 15:03.04 | dataw0lf | amount of activity |
| 15:03.06 | dataw0lf | etc |
| 15:03.12 | dataw0lf | run a benchmark |
| 15:03.16 | dataw0lf | *shrug* |
| 15:03.53 | mindjuju | k, dw, that is what i was wondering. scalabiity, is this going to be the most effective way in the future if the site makes a billion dollars and have a million clients |
| 15:04.13 | mindjuju | i guess by that time though, the code would need to be rewritten anyways to make more efficent |
| 15:04.26 | RyanE | well, you avoid an additional select when you do the update, but you lock the record (or table, depending on table type) when you do the update. |
| 15:04.32 | mindjuju | so efficent to the workload present |
| 15:05.00 | dataw0lf | I wouldn't check for things like this. |
| 15:05.01 | RyanE | so it kind of depends on the read / update ratio, but you won't know that until you have run the site for some time, and Looong before you have a million customers all doing transactions at the same time. |
| 15:05.24 | dataw0lf | checking to see if specific fields were updated or not.. meh, the overhead is minimal. |
| 15:05.29 | RyanE | The only other possibly valid reason is if you need to keep track of when something was updated. If that's the case, you wouldn't want to update the 'updated' field... |
| 15:05.39 | RyanE | right, but in most cases it's not necessary |
| 15:06.05 | dataw0lf | That's what I'm arguing. |
| 15:06.16 | dataw0lf | The overhead on doing a mass update is minimal. |
| 15:06.27 | dataw0lf | and checking to see which fields were updated isn't worth the effort. |
| 15:06.29 | RyanE | in *most* cases... |
| 15:06.45 | mindjuju | cool, that helps |
| 15:07.00 | RyanE | that's what I'm arguing as well. However, there are edge cases in which an update is considerably costly, but you won't get to those until you have many other issues to worry about. |
| 15:07.01 | dataw0lf | RyanE: Which is what I'm going to argue when I don't have the db structure in front of me |
| 15:07.33 | dataw0lf | Or the code |
| 15:07.33 | dataw0lf | etc |
| 15:08.11 | dataw0lf | You'll know when a general update is going to cost you. |
| 15:08.20 | dataw0lf | large fields |
| 15:08.20 | dataw0lf | etc |
| 15:09.09 | RyanE | when many updates happen on indices is also a bad thing. Of course, that'll only happen if you don't know how to design your tables. |
| 15:09.19 | RyanE | but it's not uncommon to have idiots doing the DB design. |
| 15:12.55 | dataw0lf | yeah. |
| 15:13.18 | dataw0lf | welcome to the wonderful world of PHP |
| 15:13.43 | dataw0lf | there should be a taser built into keyboards |
| 15:13.58 | dataw0lf | so when someone decides to learn PHP as their first language it stuns them |
| 15:14.09 | dataw0lf | and then opens up a browser to a more suitable language |
| 15:14.37 | dataw0lf | where's jsmith when you need a prototype |
| 15:15.54 | mindjuju | where is jsmith? |
| 15:15.54 | mindjuju | ~jsmith |
| 15:15.55 | jbot | jsmith is probably perpetually hungry, or the author of Asterisk: The Future of Telephony |
| 15:15.55 | mindjuju | ~seen jsmith |
| 15:15.57 | jbot | jsmith <n=jsmith@64.0.193.23> was last seen on IRC in channel #utah, 19h 30m 22s ago, saying: 'elg: Why not write a "grep" plugin for firefox :-)'. |
| 15:22.14 | *** join/#uphpu optikal (n=optikal@64.50.125.200) |
| 15:33.01 | *** join/#uphpu Alphahawk (n=Alphahaw@70.56.105.158) |
| 16:09.40 | *** join/#uphpu [-] (n=dataw0lf@66.219.227.114) |
| 16:12.20 | macnewbold | mindjuju: the difference is that a read is cheap, and many reads can happen at once, but a write is expensive and only one write (and no reads) can happen at once. So the read to pull the old values is always cheap, but the write to send the same values might not be. Plus, there are times you may need to know if something changed or not, like for the "updated" field. |
| 16:12.45 | macnewbold | but [-] and RyanE are right - unless you've got perf. problems, just doing the write is as good as any. |
| 16:30.38 | *** join/#uphpu [-] (n=dataw0lf@66.219.227.114) |
| 17:06.02 | *** part/#uphpu mindjuju (n=mindjuju@216.20.238.225) |
| 17:12.08 | *** join/#uphpu fungus (n=fungus@firebat.aros.net) |
| 17:24.21 | *** part/#uphpu fungus (n=fungus@firebat.aros.net) |
| 17:25.02 | *** join/#uphpu fungus (n=fungus@2001:470:1f00:645:20a:95ff:fec5:636a) |
| 17:26.02 | *** join/#uphpu beandog (n=sdibb@64.50.56.200) |
| 17:26.06 | beandog | whats up with the mailing list? |
| 17:26.12 | beandog | I posted a message an hour ago and it still hasnt gone through |
| 17:33.04 | *** join/#uphpu mindjuju (n=mindjuju@216.20.238.225) |
| 17:59.49 | *** join/#uphpu eco (n=eco@gnuk.net) |
| 19:08.58 | mindjuju | totally wired homes come to utah |
| 19:08.59 | mindjuju | http://news.yahoo.com/s/ap/20051004/ap_on_hi_te/digital_homes |
| 19:09.44 | mindjuju | can you imagine if somebody put up a wifi in their home without adequate protection? wardrivers able to turn on your ac to super cold |
| 19:10.03 | mindjuju | could be funny |
| 19:14.08 | fungus | Or better, disable the AC completely. In St. George too. :) |
| 19:16.00 | fungus | Hmm, without proper sanity checks, a system like that could kill the occupants. |
| 19:16.40 | mindjuju | it really could. write a little script and store it somewhere that kept telling the AC to turn off on a paticularly roasty warm day |
| 19:16.44 | RyanE | fungus: the Honeywell thermostats in that example, are 3rd-party devices, which will have sane limits and sanity checks, and will function just fine even without the Home Automation system. |
| 19:17.09 | RyanE | but yes, someone could turn the AC off. What they couldn't do is set it to invalid values. |
| 19:17.25 | mindjuju | like -10c ? |
| 19:17.32 | RyanE | but any device you put unknowingly on an open network is a disaster. |
| 19:17.51 | mindjuju | could you imagine an ac unit trying to hit -10c in the middle of the desert? |
| 19:17.57 | mindjuju | good thing for the sanity checks |
| 19:19.45 | mindjuju | could you imagine in 10 years, Symantec will have Norton Anti-Virus Home edition. with the all new refridge scanner to keep those pesky viruses away from your ice cream? |
| 19:19.49 | RyanE | most of the time, it's not that interesting for someone to remotely change your temps or turn on and off your lights, though. |
| 19:20.07 | RyanE | i.e. you can turn on and off my lights remotely, but since you get no feedback other than a text changing color, what's the point? |
| 19:20.17 | macnewbold | yeah, it's boring |
| 19:20.19 | RyanE | you can't hear me cuss at you from the internet. |
| 19:20.21 | macnewbold | unless you know he's their |
| 19:20.23 | macnewbold | there |
| 19:20.27 | macnewbold | and cussing at you |
| 19:20.31 | RyanE | and most of the time, I'm not (now that I don't work at home) |
| 19:20.49 | macnewbold | now being able to remotely flush the toilet or something, that could be cool |
| 19:20.58 | macnewbold | or turn on and off the bedroom lights |
| 19:21.01 | mindjuju | that would be cool |
| 19:21.15 | RyanE | again, pointless. I'll put a button on my website that says 'flush toilet', and you can click it to your hearts' content. |
| 19:21.22 | RyanE | without you *knowing* that it's doing it, it's pointless. |
| 19:21.37 | RyanE | oh, I have good pipes. No scalding when toilets flush |
| 19:21.58 | macnewbold | so make the toilet flush button play a .wav file for me when I click it, so I can hear it flush |
| 19:22.38 | RyanE | uh... I think the wife would frown on making audio from the bathroom available over the internet. A bit too '5th grade'. |
| 19:23.01 | RyanE | although I'm sure it'd be popular. (poopular?) |
| 19:23.03 | RyanE | :) |
| 19:39.23 | *** part/#uphpu beandog (n=sdibb@64.50.56.200) |
| 19:39.42 | macnewbold | no, it wouldn't be live audio - that's the trick... you just play a .wav file of a toilet flush |
| 19:40.00 | macnewbold | ~toilet |
| 19:40.01 | jbot | extra, extra, read all about it, toilet is not here |
| 19:40.01 | macnewbold | ~flush |
| 19:40.04 | macnewbold | ~toiletflush |
| 19:40.08 | macnewbold | ~toilet flush |
| 19:42.14 | macnewbold | ~wantnick |
| 19:42.18 | macnewbold | ~wantnick jbot |
| 19:42.18 | jbot | error: you do not have enough flags for that. (o required) |
| 19:42.28 | macnewbold | jbot, wantnick jbot |
| 19:42.28 | jbot | error: you do not have enough flags for that. (o required) |
| 19:43.00 | mindjuju | what is wantnick? |
| 19:43.25 | macnewbold | it makes jbot_ nick himself back to jbot |
| 19:43.42 | macnewbold | ~seen TimRiker |
| 19:43.46 | jbot | timriker is currently on #botpark (3h 10m 37s) #linuxfund (3h 10m 37s) #ol (3h 10m 37s) #bzflag (3h 10m 37s) #utah (3h 10m 37s) # (3h 10m 37s) #fossfund (3h 10m 37s) #classiccmp (3h 10m 37s) #elinux (3h 10m 37s) #debian-bots (3h 10m 37s) |
| 19:44.15 | mindjuju | ~flags |
| 19:47.40 | mindjuju | ~dance jbot |
| 19:52.39 | *** join/#uphpu fungus (n=fungus@2001:470:1f00:645:20a:95ff:fec5:636a) |
| 20:02.03 | *** join/#uphpu fungus (n=fungus@2001:470:1f00:645:20a:95ff:fec5:636a) |
| 20:07.18 | macnewbold | ~dance |
| 20:07.19 | jbot | ACTION becomes steve ballmer |
| 21:05.37 | mindjuju | wasn't somebody in here saying the other day how good dreamhost hosting? |