Template talk:Spellcheck Project

From Discworld MUD Wiki
Revision as of 02:03, 6 August 2009 by Taepha (Talk | contribs) (Comment on ritual capitalisation)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

TBH, I think it'd be preferable if we actually gave the spellcheck research results as well as linking to the spellcheck project - that way users can get the spell info at a glance without having to go cross-site navigating. Template:Spellcheck allows you create the standard pretty-coloured-bars effects; see Mother Harblist's Fruity Flyer for a case of it being used.

In general, I feel we should be attempting to get this site as autonomous as possible - hence, avoiding using external links as a replacement for articles or information (such as priest rituals and spellcheck data). The benefits of this are:

  • It's easier for the data to be updated and/or edited (as you don't have to worry about cross-site logins and account creation; or complete lack of control in the case of non-wiki external links).
  • You can see the relevant data at a glance, as opposed to having to navigate for it.
  • This wiki isn't degraded if the external site goes away/has an outage/gets virally infected/etc.

Of course, that's not to say that we shouldn't have any links to such sites - it would, for example, be fine to put them in the 'external links' sections in appropriate places - I just don't think we should be using links to them instead of creating articles/sections here.
--Chat 17:41, 5 August 2009 (UTC)

That would be preferable, although I've been having just a little trouble actually extracting the relevant data from the spellcheck project tables... this is giving me hope, though, since it seems to be saying that wizard spells all use a decay skillcheck with a half-life of either 36 (starting point for the "equal likelihood" message) or 41 (midpoint for the "equal likelihood" message). Do you know if the 50% success point is at the beginning of the "equal likelihood" range, or in the middle? The "am I eyeballing the right starting point" issue makes me a little uncomfortable, but I guess it doesn't make any practical difference if it's a point or two off.
RE the priest wiki--there didn't seem to be much interest in merging it properly, and the template seemed like a good stopgap, since 1)if it ever does get merged, it'll be easy to just change the template to link internally; we won't have to run around changing a bunch of links (I mean, it'd still be good to change them to normal wikilinks, but they wouldn't actually be broken in the meantime), and 2)in the meantime, the information is more easily accessible--there isn't the appearance of this huge blank area marked "Priests".
I'm also very leery of the idea of just copying everything over--it seems quite rude, since it's still active and being maintained/administered, plus I suspect we'd get a divergence problem. IMO the two good choices are either moving it all over here (I mean, as opposed to copying) in a proper merge, or leaving them separate except for crosslinks--not making a lot of duplicate articles (I'm not trying to put words in your mouth; that's just what I see as the three basic options (well, ignoring the fourth option of just doing nothing, since I suspect that would just lead to duplicate articles anyway--a red link cries out for the page to be made)).--Ilde 20:28, 5 August 2009 (UTC)
Having looked at the mudlib, I strongly suspect the information from here is incorrect - the mudlib treats all fixed skillchecks as either linear or decay-based (see taskmaster for more details). The only way the data from that link could be correct is if the taskmaster's comparison system was specifically overridden with something else for the purposes of spell stages. I doubt someone would go through that degree of effort to produce something which is mostly similar to the built-in decay skillcheck.
As for the half-life and cutoffs, well, the cutoff is clearly entirely variable based on the spell stage. The half-life has probably been specifically set for each spell stage too, however, it's also possible that various cres just left some half-lives at their default, which is 6 * sqrt(cutoff).
Regarding the categories, there's no data in the mudlib on exactly which chance of success they correspond to, so we're just going to have to guess. I've been taking 'most certainly fail' as 0% success, and all subsequent categories as 10% bands.
My worry with the priestwiki stuff is that it's going to stop people from actually creating the relevant articles - they'll see a bluelink and assume that the article's there. It might be better to create a stub and just put the link to priestwiki in it instead. While there'll still be a bluelink in that case, the soft-redirect means that people will see the article needs filling out, and the stub template will auto-insert the relevant article into the list that need attention.
Agree we don't want to 'just copy' things from external sites without permission, as this will doubtless annoy whoever has been carefully collating the data. Ideally, we want to persuade the owners of the relevant sites to transfer their data here; of course, that is going to be difficult while this site gets relatively low traffic.
--Chat 21:38, 5 August 2009 (UTC)
I'm not sure at what level you are rating websites in terms of their traffic, but just as an aside the wiki has around 400 unique visitors per day, according to webalizer. 21:48, 5 August 2009 (UTC)
Hmmm... well, I know it doesn't match the linked examples(!) but I'd assumed it was just an expression of a decay-based skillcheck. (I read the taskmaster page, and it seemed to fit, but I admit I didn't really work out when you'd see the different levels with a half-life of 36/41; just noted that the distance between levels increased as you went on and that it seemed to work with 36 and 72, or 41 and 82.)
It does (somewhat) look as though the half-life for all the wizard spells is the same, though: if you look at any particular spell (or skill, in fact--like the often-used channeling) then it appears as though the range between "most certainly fail" and "most certainly succeed" is the same for each. (You do have to look at the numbers, though, not the apparent width of each range, since the first few columns are inexplicably narrower.)
That's a good point about it potentially discouraging people from making the articles... I think I could just use the "what links here" for the template to backtrack and see what needs making. Just offhand I don't think there are all that many articles linked to. --Ilde 23:14, 5 August 2009 (UTC)
There we go... the only pages still using the template are the soft redirects. Like... Pishe! --Ilde 01:08, 6 August 2009 (UTC)
Good good! We should probably get a consensus oh weather or not we should fully capitalise ritutals, they're not like spells. Rehevkor 04:23, 6 August 2009 (UTC)
I vote yes, because 1)it will look more consistent than if we have spells capitalized and rituals not, and 2)they're capitalized in the help files, even if they aren't everywhere. --Ilde 04:49, 6 August 2009 (UTC)
I vote no. The rituals themselves are not captalised (in 'rituals'). Capitalisation in helpfiles is not completely consistent (see 'help remember place', 'help soothe'). I think, within the helpfiles, it's simply a case of whoever wrote them applying their own preference, which does not appear to be consistent with everything else. I'm not completely opposed to capitalisation in this case, as it's so variable in the game; my preference is primarily based on the uncapitalised names in 'rituals'.