Coding policy/ some Q&A..

[carolynwhite2]

Dot posed me some good questions this morning, and with her 
permission I am re-posting the Q&A as I thought other people might be 
interested as well.

PS Loved the naked wizards, Laurasia! And BTW, I got the Avery post 
number wrong a few posts back - it was 34911, not 34811...sorry.

*******************************************

Dot:
My first thread confused me a little (36401, All things green), it's 
a list of references to the word `green', canon quoted, but not 
discussed, and though there are a lot of posts the list is far from 
exhaustive.  I coded it simply as `green', but having gone through to 
the end of the thread there's absolutely no discussion at all, so I'm 
now tempted to reject the lot, what should I do?

Carolyn:
This is a LOON-type thread, a listing of all mentions of 'green' in 
the first four books. (LOON = League of Obsessed Nitpickers..). 
Although there is no discussion, I would keep it on a 'for-the-record-
 basis, because I know this is just the kind of the thing people like 
to look up, and who knows, it may eventually be relevant to some 
future plot development. In past coding, I have also kept various 
long lists of biblical references, vampire-references and so on for 
the same reason.

Dot:
Secondly – how do you deal in general with threads that start before 
the batch of posts you're looking at, and/or end after?  For example 
there's part of a discussion on how Hogwarts pays for itself which is 
interesting, but the more interesting and well-thought-out posts are 
later on, after the batch that I'm looking at.  Should I reject the 
two that come within my remit on the basis that there's little canon 
in them and better stuff which repeats those arguments later on, and 
write notes at the bottom in the hope that whoever gets those posts 
checks to see why I excluded them?

Carolyn:
This depends on how much time you want to spend on it. It is great to 
look up and down thread, but sometimes those threads are very long. 
After you have done quite a bit of coding, I find a compromise 
approach usually evolves. There are themes which come up a lot and 
you are personally quite *certain* that you have saved numerous, and 
possibly better posts on a particular subject. This prompts you to 
hit that reject button rather tartly. 
 
OTOH, if it feels like a newish topic or at least an original or 
newish take on an old topic and you have landed in the middle of it, 
then yes, it is worth taking a bit of time to check up thread, 
particularly if there are big-hitter posters engaged in one of their 
entertaining debates. Best to keep the thread in it's entirety then - 
although, even so, exclude 'I agree' or really silly OT/adds totally 
nothing new posts. Hard to legislate...
 
On balance, I would err in favour of saving rather than rejecting if 
you can't decide. We will be going through a second-edit stage, and 
can reject unnecessary duplication then, if we wish.

Dot:
Thirdly, how much coding is enough?  Should every character that is 
mentioned in the post be checked?  Every theory acronym?  Or only 
those that are significantly discussed?

Carolyn:
I try to go for elegant minimalism, to capture the essence of the 
thread. However, that said, yes, every acronym mentioned should be 
ticked (unless it is just a passing wave from one dinghy to another), 
and the key characters mentioned. For example, if the thread 
discusses Snape's teaching style, I would not necessarily tick Harry, 
Neville and so on, unless there was a lot of talk about how they felt 
about it. 
 
But on the big theory threads we are now encountering, it is really 
not uncommon to be ticking 20-30 categories sometimes.

Dot:
Fourthly, multiple posts where some parts are FAQ/adds nothing new, 
but other parts are worthy of coding - should only the interesting 
bits be coded, or should the whole post (including the uniteresting 
bits) get codes too? For example post 36422 is a TBAY summary for 
newbies of a lot of theories that were washing around at the time - 
there's a lot that's not new at all.  However, there are a couple of 
short responses to posts hidden in there, so what should get coded?  
(I ended up coding the lot as it's quite a nice introduction, but I 
don't know if that's right.)

Carolyn:
Oo..nice post that, very useful. A post like this I would tick 
everything relevant in sight. The problem comes on multi-posts like 
CatLady's where some things should be coded, others are potentially 
reject candidates, but you can't mix reject and non-reject codes in 
the same post, so you end up just coding some of it.
 
Eventually, the way we plan to deal with multi-posts is to have the 
parts relevant to codes highlighted when people do a search and 
retrieve on the catalogue, so the non-relevant parts don't stand out.
 
So, the short answer is, if a post has even only one small part that 
is worth keeping, code that and ignore the rest.

Dot:
Fifthly, I'm worried about how many posts I should be rejecting (I 
want to reject more).  Where there's a small paragraph with a 
question that is repeated in the replies I'm rejecting them.  But 
then  there's (for example) post 36430 has a paragraph about Rita 
Skeeter (which I'd count as 'adds nothing new' in the context of the 
thread) but then some speculation about Harry being a snake animagus, 
which we now know is not possible as JKR said none of the trio will 
be animagi.  Should that go under 'mistakes/perpetrating mistakes'?

Carolyn:
Yes, I'd like to keep our reject rate as high as possible, consistent 
with fairness. It has now slipped down to 56% from our initial high 
of 70%, alas.
 
Rejecting initial questions that get repeated later in replies is 
good, and you should always do this if possible.
 
What constitutes a 'mistake' is a tricky one, however. You may have 
seen earlier discussion on the catalogue site as to whether we should 
now reject all previous posts about the Weasley's ages, as JKR has 
now 'confirmed' them on her site (allegedly). We decided not to do 
that, on the grounds that this would remove too many posts that 
should be kept for posterity (the Florence/Bellatrix confusion is 
another example). 
 
The example you quote is also difficult [36430], as although the trio 
are not going to be animagi, apparently, it is speculating in a 
general way as to what might happen if a person in animagi form got 
killed or eaten by an animal. I would therefore keep this one and 
code to animagi, and probably also Rita Skeeter and Hermione.
 
Probably no one should remind the protagonists on the current 
Marietta thread as to this further instance of Hermione's moral 
dubiousness !