Perl 6 - the future is here, just unevenly distributed

IRC log for #rosettacode, 2011-06-24

| Channels | #rosettacode index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:00 mwn3d_phone left #rosettacode
00:06 soniakeys joined #rosettacode
00:35 FireFly left #rosettacode
00:53 mwn3d_phone joined #rosettacode
01:33 dagnyscott joined #rosettacode
01:59 kpreid left #rosettacode
02:45 kpreid joined #rosettacode
03:26 mikemol Coderjoe: If I attempt an update, and it fails, and then I insert, then I may have two processes attempt an update, fail, and attempt inserts.
03:27 mikemol insert, update if it fails won't race. OTOH, I could update, insert if fails, update if insert fails.
03:27 mikemol That would reduce in less overhead from exceptions, overall. (Majority of cases won't see exception, as opposed to the current case of majority of cases seeing exceptions)
03:29 mikemol Idea I had a few minutes ago...it'd be really useful if tasks specified expected output for expected input (even as a separate page or resource, if necessary), and that could be uses as the basis for unit tests for code examples for tasks which produce output.
04:11 dagnyscott left #rosettacode
04:35 r4m left #rosettacode
08:39 mwn3d_phone left #rosettacode
09:03 slavik left #rosettacode
09:58 aLeSD joined #rosettacode
10:49 mwn3d_phone joined #rosettacode
10:55 soniakeys left #rosettacode
11:15 aLeSD left #rosettacode
11:15 aLeSD joined #rosettacode
11:54 FireFly joined #rosettacode
12:45 dagnyscott joined #rosettacode
13:04 Jerin_Philip joined #rosettacode
14:03 Jerin_Philip is now known as kutti-chat-an
14:03 kutti-chat-an left #rosettacode
14:49 dagnyscott1 joined #rosettacode
14:55 dagnyscott left #rosettacode
14:55 dagnyscott1 left #rosettacode
14:55 dagnyscott joined #rosettacode
15:18 * mikemol needs to not use GNU-screen commands for navigating in irssi.
15:18 mikemol So, earlier this week, I was contacted by the owner of CodeEval.com. He wondered if there were a way CE and RC could work together.
15:18 fedaykin "CodeEval - Evaluations Made Easy"
15:19 mikemol Last night, I had an idea. I like it enough that I expect I'll implement it anyway, and open its use up to anyone who wants to take ownership of compiling arbitrary data and sending back the output.
15:22 mikemol In the handler for the <lang> tag, I can take the tag's content, sign it, use the digest signature as a local cache key and transaction identifier, POST the code and signature to CE.
15:22 kpreid left #rosettacode
15:23 mikemol When CE's done chewing on the data, they can POST the data (including the signature I sent them, and their own signature) back to my server.
15:23 mikemol My server would save the output in a lookup table keyed to the signature I'd used in sending the data to them.
15:24 mikemol When the <lang> tag is processed, prior to sending the data to CE, I would check the lookup table to see if the data was already there. If it's already there, I can use that, and skip the POST.
15:24 slavik joined #rosettacode
15:25 mikemol Since <lang> can be hit many hundreds of times in a single page, I would need to limit the number of queries-out I do per page render. Probably leave it at something like 2-3.
15:26 mikemol Where's BenBE when you need him? This'd interest him, too, as I recall he wanted to try setting up centralized highlighting.
15:27 mikemol (Obviously, in any case, I'd need to sanitize data before I passed it out to a browser)
15:27 Hypftier You can do that as a background job too, not necessary to tie it to page edits or displays. Just have a queue and slowly work on that, right?
15:27 mikemol Hypftier: Queuing of the item would have to be done in the <lang> processor, at the very least. But, otherwise, sure.
15:28 mikemol In fact, that's probably the better approach, for a variety of reasons.
15:35 mwn3d_phone So wait....what would we be getting from CE then?
15:35 mwn3d_phone Code output?
15:45 mikemol Yeah
16:13 kpreid joined #rosettacode
17:46 soniakeys joined #rosettacode
18:17 * mikemol jawdrops
18:17 mikemol We've got 529 tasks?
18:18 soniakeys and 386 languages
18:19 mikemol 238,969 page views per month.
18:20 sorear is that a lot?
18:23 mwn3d_phone 238k page views a month? I thought analytics had us at like 3500 a day
18:29 mikemol mwn3d_phone: Visits, yes. Pageviews, no.
18:30 mwn3d_phone Ohhhhhhhh. Tricky tricky.
18:30 mikemol sorear: For a major commercial interest, no. For something I still keep thinking of as unknown, I think so.
18:31 mikemol That said, it no longer surprises me when friends and family members say, "hey, I ran into someone who knows about Rosetta Code!"
20:02 benbe joined #rosettacode
20:04 benbe2 joined #rosettacode
20:05 benbe2 To answer the question of where I was:
20:06 benbe2 Buying a new router since the previous one died of old age yesterday.
20:06 benbe left #rosettacode
20:06 benbe2 left #rosettacode
20:06 benbe joined #rosettacode
20:09 benbe left #rosettacode
20:09 benbe2 joined #rosettacode
20:09 benbe2 left #rosettacode
20:09 benbe2 joined #rosettacode
20:18 benbe2 left #rosettacode
20:23 soniakeys left #rosettacode
21:01 FireFly left #rosettacode
21:05 dagnyscott left #rosettacode
23:18 TimToady yay, Perl 6 hits the 300 mark!
23:25 realazthat_ joined #rosettacode
23:28 realzies left #rosettacode
23:33 mwn3d_phone TimToady: o/
23:35 mwn3d_phone (Its a guy soliciting a high five in case you haven't seen it before)
23:46 TimToady oh, we just think it's waving on #perl6
23:46 TimToady \o
23:47 TimToady we use \o/ for celebrating though.... :)
23:52 mwn3d_phone My friends and I use \o/ for self-celebrating. We usually use o/ for especially good insults.
23:53 TimToady well, we don't know how to insult each other very well in Perl culture.  :P
23:55 TimToady and as for "self-celebration", we are all completely ego-less (yeah, sure... :)

| Channels | #rosettacode index | Today | | Search | Google Search | Plain-Text | summary