Perl 6 - the future is here, just unevenly distributed

IRC log for #rosettacode, 2011-08-24

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

All times shown according to UTC.

Time Nick Message
00:02 kingxerxes_ joined #rosettacode
03:38 BenBE joined #rosettacode
09:51 p6eval joined #rosettacode
10:43 FireFly joined #rosettacode
11:26 kingxerxes joined #rosettacode
12:52 dagnyscott joined #rosettacode
17:50 mikemol http://soniacodes.wordpress.com/2011/02/06/a-month-of-rc/
17:50 fedaykin "A month of RC | Sonia Codes" http://rldn.net/9z0
17:51 mikemol mwn3d_phone: Probably a good basis for some categorization, at least.
17:58 mwn3d_phone Seems ok but I'm bitter about categories now. I want tags to work.
17:58 mikemol Hehe
17:58 mikemol Good man. :)
17:59 mwn3d_phone If we want to do categories again we should undo all of the categorization we have
18:00 mwn3d_phone But if we do that then it'd probably be better to keep some category pages around
18:00 mwn3d_phone But then we'd be planning and maybe we'd take away from "natural" categories
18:00 mwn3d_phone Blech....what a mess
18:00 * mikemol nods
18:02 mwn3d_phone The SMW projects I really want to work are tags in place of categories, transcluding separate example pages on to task pages, and a form for adding an example to a task
18:03 mwn3d_phone I wish we had an SMW equivalent of BenBE to help us use things correctly
18:21 mwn3d_phone1 joined #rosettacode
18:39 kingxerxes joined #rosettacode
19:57 dagnyscott joined #rosettacode
19:59 Kumul joined #rosettacode
19:59 Kumul ello, is there a rosettacode repo somewhere?
19:59 mikemol No
20:00 mikemol I've actually been thinking about that lately, wishing RC were backed by, e.g. git.
20:00 mikemol It would be royally non-trivial to hack that into MW, though.
20:07 Kumul i would prefer a git web interface to MW, since its made to appeal developers
20:11 Kumul oh well, i'm not qualified to offer opinion on this matter. but thanks, was just curious.
20:11 mikemol Sorry..difficult to respond. At work right now.
20:11 Kumul k :)
20:11 mikemol I'd prefer git, too, simply for the benefits it'd give for backup and distributions.
20:59 eel2 joined #rosettacode
21:12 kingxerxes joined #rosettacode
21:17 Yaron1 joined #rosettacode
21:19 Yaron Anyone need help with SMW?
21:25 Yaron (Semantic MediaWiki, that is.)
21:28 mwn3d_phone Yaron: we want to get SMW to help us organize tasks
21:28 mwn3d_phone At the very least we want to be able to add tags to tasks
21:29 Yaron Alright, cool.
21:29 mwn3d_phone We would tag them with programming concepts of some sort
21:29 mwn3d_phone Another thing that would be nice is to have SMW handle code examples for us.
21:29 Yaron Alright.
21:30 mwn3d_phone What I keep bringing up but can't tell if anone else wants it is that we would have examples on separate pages tagged with a language name and a task name
21:30 mwn3d_phone Then we would somehow transclude their text on to the task pages in alphabetical order by language
21:31 mwn3d_phone It would be nice to have a form for adding an example to a page (give the language name and a block of text)
21:31 Yaron Well, it looks like your wiki already makes extensive use of SMW, and of Semantic Forms...
21:31 mwn3d_phone Someone started that but I don't know how well it works
21:32 mwn3d_phone Forms not so much I don't think. We have them. I don't think they work well and I don't think people use them.
21:32 Yaron Ah.
21:32 mwn3d_phone I also tried to get another way of organizing tasks going which would go alongside the tags here: http://rosettacode.org/wiki/Rosetta_Code:Village_Pump/SMW_Examples_by_language_and_concept
21:32 fedaykin "Rosetta Code:Village Pump/SMW Examples by language and concept - Rosetta Code" http://rldn.net/2Yd
21:33 mwn3d_phone I wasn't able to convince mikemol though and no one else joined in
21:33 Yaron Well, it sounds like your problem might be a social/political one, rather than a technical one...
21:34 mwn3d_phone (Not that you were trying to get in the way, mikemol. You were just the only person that said anything.)
21:34 mwn3d_phone Well also I don't know how to get SMW to do these kinds of things in a way that can replace wiki categories
21:35 mwn3d_phone We would want tag searching and probably a listing of tags on each task page
21:35 mwn3d_phone I couldn't get the #ask query right to do that last part
21:35 Yaron Is the failed #ask query still there on the wiki?
21:36 mwn3d_phone I don't think so. Basically I just wanted to list all values for a particular tag on "this" page.
21:36 mwn3d_phone It seems like something easy for someone who knows what there doing but I don't and the documentation for SMW didn't help me much
21:37 Yaron Sorry, I don't know what you mean by "tag", "task" and "this".
21:38 mwn3d_phone This: the page that the query is on
21:38 mwn3d_phone Tag: the concept we want to use to replace categories
21:39 mwn3d_phone Task: the basic page on RC
21:39 mwn3d_phone Are you a user of the site?
21:39 Yaron No.
21:39 mwn3d_phone Sorry I thought you were
21:39 Yaron So does each tag value have its own page?
21:40 mwn3d_phone I'm not sure if that would happen or not
21:40 mwn3d_phone Brb...gotta finish my commute in the car
21:40 Yaron OMG.
22:01 mikemol mwn3d_phone: I flagged that VP discussion as something I wanted to get more involved with, but I got really busy, IRC.
22:01 mikemol *IIRC
22:01 mikemol Yaron: Here from the Twitter of FB request, I assume?
22:02 mikemol s/of/or/
22:02 mikemol My god, I don't know what's rong with me.
22:02 mikemol -.-
22:02 Yaron Yeah.
22:02 mikemol Yaron: Many, many thanks for responding.
22:02 * mikemol goes over discussion backlog
22:05 mikemol Ok, quick overview of the basic site structure.
22:06 mikemol We have "tasks" which are generally requests for code; "show how to do this" or "demonstrate that". Each task can be implemented any number of times using any combination of tools.
22:06 mikemol Right now, the primary class of tools demonstrated on RC are langauges. You might see the 99 Bottles of Beer problem demonstrated in 300 different languages.
22:07 mikemol To a limited extent, there are some other differentiations. For [[GUI/Create a window]], you might see a dozen langauges, but the C language might be combined with two or three different GUI toolkits.
22:07 fedaykin http://rosettacode.org/wiki/GUI/Create_a_window (Doesn't exist.)
22:07 mikemol er.
22:08 mikemol [[Window creation]]
22:08 fedaykin http://rosettacode.org/wiki/Window_creation
22:08 mikemol MediaWiki categories are used extensively.
22:09 mikemol We have a category for each langauge. We have a category for each library. We have a category containing all language categories as a subcategory. We have several categories *per* language, as subcategories of the category named after the language.
22:09 mikemol If that sounds confusing, that's because it probably is.
22:09 mikemol Hm
22:10 mikemol If you look at [[Category:C]], you'll see three subcategories.
22:10 fedaykin http://rosettacode.org/wiki/Category:C
22:10 mikemol If you look at [[Category:Programming Languages]], you'll currently see 393 subcategories.
22:10 fedaykin http://rosettacode.org/wiki/Category:Programming_Languages
22:11 mikemol If you scroll to the bottom of [[Category:C]], you'll see that it's a subcategory of several other categories, most of which relate to tagging the C language as having a variety of properties.
22:11 fedaykin http://rosettacode.org/wiki/Category:C
22:12 mikemol (For example, [[Category:Parameter passing/By value]])
22:12 fedaykin http://rosettacode.org/wiki/Category:Parameter_passing/By_value
22:12 Yaron Yikes.
22:12 mikemol And that's just looking at things from the languages' perspective.
22:13 Yaron Right.
22:13 mikemol From the perspective of task pages, let's take a look at [[Synchronous concurrency]]
22:13 fedaykin http://rosettacode.org/wiki/Synchronous_concurrency
22:14 mikemol First thing that's obvious: It's a task; it asks for code.
22:14 mikemol Looks like, at the moment, it's implemented in 31 languages.
22:14 Yaron I think I have enough information now.
22:15 Yaron I think the basic approach should be to create a "Task" template, that has fields, which themselves are stored with semantic properties, for each "facet" of data around a task.
22:16 Yaron In this case: language(s), library/ies, ...
22:17 Yaron language(s) in which it can't be implemented...
22:17 mikemol afk a second. Go ahead and elaborate, I'll read in a minute or two.
22:17 Yaron Alright.
22:18 Yaron ...and task type (games, encryption, etc. etc.)
22:18 Yaron So I see four fields/properties for this template.
22:18 Yaron ...and then you should have a form for tasks as well.
22:18 Yaron I would use this page to do it: http://rosettacode.org/wiki/Special:CreateClass
22:18 fedaykin "Permission error - Rosetta Code"
22:19 Yaron And then, instead of pages like "Category:C", you should just have pages called "C".
22:21 Yaron ...which use a template called "Programming language", which in turn contain one or more #ask queries, like {{#ask:[[Category:Programming tasks]][[Has programming language::{{PAGENAME}}]]}}
22:21 fedaykin http://rosettacode.org/wiki/Category:Programming_tasks (Doesn't exist.)  http://rosettacode.org/wiki/Has_programming_language::{{PAGENAME}}
22:22 mikemol One other thing I would like to be able to do: Get rid of task pages as they exist, and populate them using semantic inline queries.
22:22 mikemol Coderjoe built a template that we thought would do that, but it didn't work for some reason.
22:23 mikemol We were using [[Hough transform]] as a testbed for it.
22:23 fedaykin http://rosettacode.org/wiki/Hough_transform
22:24 mikemol Looks like the old discussion around that was at [[Rosetta Code:Village Pump/Semantic MediaWiki/Semantics]]
22:24 fedaykin http://rosettacode.org/wiki/Rosetta_Code:Village_Pump/Semantic_MediaWiki/Semantics
22:25 * mikemol digs through wiki history to find the template in question
22:25 Coderjoe the problem was that SMW was taking the properties of the transcluded pages and applying them to the page they were transcluded into
22:25 Coderjoe and I was playing with User:Coderjoe/Sandbox
22:25 mikemol [[User:Coderjoe/Sandbox]]
22:25 fedaykin http://rosettacode.org/wiki/User:Coderjoe/Sandbox
22:26 mwn3d_phone joined #rosettacode
22:26 mikemol mwn3d_phone: You'll want to read the backlog
22:26 Yaron mikemol - you want to get rid of task pages?
22:27 Yaron So where would the information be stored?
22:27 mikemol Yaron: As static pages, yeah. If possible, I'd like to have each example as its own page, and transclude those.
22:27 mikemol If that can be done, then so can a page which transcludes all examples done in a particular langauge, or done with a particular library.
22:28 mikemol I get and overhead a *lot* of wishes for cookbook layouts and layouts which omit languages a viewer isn't interested in.
22:28 Yaron Oh, I see.
22:28 mikemol If that could be done with transclusion, that would be epic. (And a royal abuse of MW and SMW, I suspect, but, well...)
22:29 Yaron Well, that's a tough problem.
22:29 * mikemol nods
22:29 mikemol Coderjoe's template almost works: http://rosettacode.org/mw/index.php?title=User:Coderjoe/Sandbox&oldid=93208
22:29 fedaykin "User:Coderjoe/Sandbox - Rosetta Code" http://rldn.net/AApO
22:29 mikemol It looks like it works fine there, but, IIRC, if it were moved to [[Hough transform]], it breaks.
22:29 Yaron There would still be semantic information that relates to the task as a whole, not to task/language combinations - like the task type.
22:29 fedaykin http://rosettacode.org/wiki/Hough_transform
22:30 mikemol Sure. I'd be happy with leaving the task description more or less where it is, and replacing the mass of the body with inline transclusion.
22:31 mikemol s/looks like it works fine/mostly looks like it works/
22:31 Coderjoe mikemol: no
22:31 mikemol Looks like there are some problems with the =={{header}}== usages.
22:31 Coderjoe it transcludes what it supposed to, but after pulling it in, look at the properties for that sandbox page
22:31 mikemol Coderjoe: Understood.
22:32 Coderjoe every property set in a transcluded page is applied to the sandbox page, too
22:32 Coderjoe so it winds up trying to transclude itself
22:32 mikemol So the problem is that some properties need to not get past a page transclusion.
22:32 Yaron Well, if you put a template call like {{Task for language}} at the top of each task/language combination page, it could work.
22:34 Coderjoe http://rosettacode.org/wiki/Special:Browse/User:Coderjoe-2FSandbox
22:34 fedaykin "Browse wiki - Rosetta Code" http://rldn.net/DfG
22:34 Yaron You can handle properties getting transcluded, or not getting transcluded, by using the <includeonly> and <noinclude> tags - I don't remember how, exactly, but it's one or the other of those that allows such a thing.
22:35 Coderjoe Yaron: that breaks the header templates then
22:35 mikemol <noinclude> would prevent transclusion. We use that for template documentation.
22:35 Yaron What header templates?
22:35 mikemol Coderjoe: I think [[Template:Header]] would need to get replaced as a mechanism, then.
22:35 fedaykin http://rosettacode.org/wiki/Template:Header
22:35 Yaron Anyway, you don't have to put the whole page text in <noinclude> or <includeonly> - just the property calls.
22:36 Coderjoe if you use noinclude, you have to put the properties in each page directly, rather than being in a template
22:36 * mikemol nods
22:36 mikemol I was just thinking about this.
22:36 Coderjoe which means all kinds of additional stuff people need to remember to add
22:37 mikemol Yeah, it's a hard problem.
22:37 Coderjoe I wanted to fix the bug, not work around it.
22:37 mikemol It's also getting really, really fuzzy.
22:37 Coderjoe but I never got a working MW+SMW testing environment to work in
22:37 Yaron Meh. I would say, just stick with having one semantic template, per main task.
22:38 Yaron It's simpler, conceptually.
22:38 Coderjoe because, if you ask me, the SMW #ask query should not allow properties to be set on the page with the query
22:38 mikemol Coderjoe: chill
22:39 mikemol A single template is a good starting place.
22:40 mikemol It should build some experience, and help those active on the wiki get familiar with how the SMW engine works.
22:40 mikemol Now if only MW supported redirecting category pages...
22:41 Coderjoe but the SMW engine is broken in a way that this "template" pollutes the properties
22:41 kpreid joined #rosettacode
22:41 Coderjoe you can use this to generate a list of pages. that works. but transclusion has issues
22:42 mikemol So we don't use transclusion yet, until we've had a change to get a proper fix done.
22:43 mikemol Coderjoe: I can't get you a copy of RC's existing db, but I can help you set up SMW. dagny and I set up an install on her system a couple weeks, ago, it took less than an hour.
22:43 Coderjoe i wasn't expecting the DB.
22:49 mikemol Yaron: We'll be poking more at this as we have time. It'd be greatly appreciated if you could lurk indefinitely. (If you don't mind, of course. :) )
22:50 Yaron What do you mean, indefinitely?
22:52 mikemol An undefined amount of time. Could be merely two more seconds; your choice. ;)
22:52 Yaron Alright - well, then, I'll definitely stay indefinitely. :)
22:53 mikemol Point is, I don't know how quickly we'll be able to act on any of this. It'll depend on who has free time. It'd be nice to have you around to poke on occasion, but I can't say how soon you might be poked. :)
22:53 Yaron Well, if I or someone else is around, you or others can always go to the #semantic-mediawiki channel, and ask there.
22:53 Yaron (isn't around)
22:54 mikemol True.
22:54 mikemol I used to hang out in there. Not sure why it dropped off my autojoin.
22:54 mikemol It can be tricky describing the site's natural ontology, though.
22:54 Yaron Indeed. :)
22:55 mikemol Yeesh. I only learne the word 'ontology' because I needed to get SMW added to RC. Need drives a *lot* of learning.
22:55 mikemol *learned
23:20 mwn3d_phone mikemol: you can redirect category pages I thought. You just can't redirect the members to that new categor.
23:51 Coderjoe joined #rosettacode

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