We thought it would be helpful to gather up the
issues that we have discussed and put them in one message for easy
reference. We've also tinkered with the database a bit and refined
how we'll handle keyword issues, so I'll explain how that has
changed.


WHAT TO INCLUDE?

Our goal, of course, is to produce a catalogue of coherent and canon-
based posts on every conceivable canon issue that has been
discussed, and particularly posts and threads that someone will
eventually use to write or update a FAQ. So we're looking for posts
that are interesting, amusing and/or well-reasoned, and that have
good analysis and good form. No theory is too outlandish to be
included in the database. Let the FAQ editor decide whether "Lupin
was an unpaid auror" makes it into the FAQ.

What posts should we *not* include in the catalogue? We should not
include posts that make no canon point (OT, LOL!, me too!), posts
that have atrocious form (spelling, grammar, etc.), extremely short
posts, posts that merely repeat what others have said much better in
the same thread, and gibberish.

That said, if you run across a truly awful thread and wind up citing
no posts at all from that thread, please do go back and pick the
best post from that discussion so that we won't miss entire threads.

What constitutes a thread? Well, that gets a bit tricky. You might
run across a single post that drew no responses, or a new post that
drew only one response, and you might find that nothing in these
brief threads is worth preserving. This can happen when someone's
message does discuss canon but really is more of a statement of
personal preference ("I sure hope Sirius doesn't die!") or when the
post that started the thread was just not very good. In these
instances, use your judgment about whether to jot down a post from
the thread. The idea here is that if there is some value in making
sure that the thread is not lost and gone forever, then you should
select a post from the thread. Otherwise, don't worry about it.

Those reviewing messages should review all types of messages except
for FILKs and ADMINs, including SHIP, FF and TBAY.

HOW TO CHOOSE KEYWORDS

When deciding what keywords to use to represent a post, you should
try to keep in mind how the completed document is likely to be
used. It is inevitable that FAQ editors and other users of the
finished product will have to enter multiple keywords in order to
find posts that suit their purposes. It is also a given that the
users will be reading the contents cell to learn what the posts are
about. The function of the keywords is to allow users to narrow
down their search; it is not to give a full description of the
content of the post. That is what the "content" cell is for.

So the key here is to keep the keywords simple. Very simple. Use
them to designate the subject of the post on the broadest level -- 
by character name, by magical item, by overall concept –- and let
the content cell serve to elaborate on the more nuanced elaborations
and distinctions.

We have allowed only three cells for keyword designation. Sometimes
you may find that this is just not enough. When this happens,
insert a new row for further description of the post. Be sure to
remember to give the new row the same message number! Even though
the message numbers of the two rows will be identical, Excel will
still not recognize them as the same, and it will not be able to
combine the keywords from both rows when it sorts. Therefore, try
to keep related keywords on the same line. A combined post that
dealt in a substantive manner with both the relationship between
Snape and Dumbledore and the relative merits of the H/H and H/G
ship, for example, might be given three separate rows in Excel,
looking something like this:

30,489; Snape; Dumbledore
30,489; SHIP; Harry; Hermione
30,489; SHIP; Harry; Ginny

This layout would ensure that a SnapeFAQ editor, a SHIPFAQ editor,
the editor of the Harry FAQ and the editor of a Ginny FAQ would all
be able to call up this post by using sensible and obvious
combinations of keywords.


CONSISTENCY ISSUES

Here are the general guidelines for selecting keywords.

1. Character names – all students go by first name, all adults
go by last, except for where this could become misleading or
confusing (Weasleys, Potters).

So it's Draco, Neville, Harry and Ron. Their fathers are Malfoy,
Longbottom, James and Arthur. Their mothers are Narcissa, Mrs.
Longbottom, Lily and Molly. Common sense should hold sway here, of
course. Obviously there is a Crouch Sr and a Crouch Jr. A post
about teenaged Riddle should be keyed to "Voldemort."

2. Animals and magical creatures should be referred to by name
(Erroll, Winky) unless discussed generically (owls, house elves).

3. Spells should be referred to in some logical fashion. Defer
to what has already been entered in the keyword database. If the
database already contains "Summoning Charm" and you prefer "Accio,"
be flexible and go with "Summoning Charm."

4. Please key contests/puzzles and results to the
keyword "Contest."

5. Posts dealing with romantic or sexual relationships between
characters should carry the keyword "SHIP" even if the poster forgot
to use the prefix on the list itself. Do not attempt to convey
ships by means of any other keyword. Instead, leave that to the
content field. A LOLLIPOPS post, for example, might have
keywords "Snape," "Lily," "James," "SHIP," and (if it were a TBAY
post) "TBAY." Do not use the slashed shipping shorthand (H/H, R/H,
etc.) as keywords.

6. For SHIP, TBAY and FF, please use the prefix and other
keywords depending on canon point. Please do not include separate
keywords for acronyms; we have over 200 different acronyms in use on
the list. Instead, use appropriate keywords to describe the canon
point of the post, using the comments field for the acronym.
Remember, garden variety acronym posts are not TBAY and so do not
require the prefix.

7. You might encounter SHIP posts and TBAY posts in which the
prefix was used improperly, either because the poster neglected to
use the appropriate prefix or because s/he used it when it was not
necessary. Go ahead and rectify these errors when you catalogue. A
post which makes reference to an established theory by means of
shorthand (ie, Fourth Man, Big Bang, MAGIC DISHWASHER) and does not
elucidate the meaning of that theory in the post should still carry
a TBAY designation, even if the poster did not write in a "TBAY
style."

8. For posts which deal with literary analysis, do your best to
describe the subject matter of the post (Foreshadowing, Deus ex
Machina, Characterization), but go with what is already in the
database whenever you can.


ADDING KEYWORDS TO THE DATABASE

As we are rapidly approaching 200 keywords, try not to duplicate
concepts in the keywords database. When sitting down to do a day's
cataloguing, make sure that you have the most recent version of the
database before you begin. When confronted with a new concept or
subject, look over the list first to make sure that someone has not
already entered a word that would suit your purposes. There is no
need to get too finicky about distinctions in connotation. If there
is already an entry for "Pride," for example, then you probably do
not need to add a new keyword for "Stubborness." The two words do
not really have the same meaning at all, but they are close enough
in how they are generally used on the list to refer to the same type
of post.

Similarly, try to keep a lookout for antonyms (Age and Youth,
Responsibility and Irresponsibility). While there are a few
exceptions, generally speaking a single keyword will suffice for
both a concept and its antithesis.

If keeping all this in mind you still find nothing in the current
database that suits, then come up with a new keyword that you feel
best expresses the concept you are attempting to define. Do try to
keep your keywords relatively broad in scope. Remember that we are
using character names as keywords, but we are not adding character
names to the database.

Before adding any new keywords to the database, be sure to refresh
the database and give it one last scan to make sure that no one has
recently added any synonyms for the concepts that you wished to
define. If someone has beaten you to the punch on the keywords,
then please go back to your catalogue and make the appropriate
changes. "Search and Replace" is your best friend here.

If, however, there are still no keywords covering the concepts you
have defined in the database, then go ahead and add your new
keywords. It is often useful to give an explanatory definition in
parantheses after the word itself (ie, "Apparate [includes
Disapparating]"). Remember that what seems an obvious usage to you
may not be at all obvious to others; try to make your meaning as
clear as you can when you add new words to the database.

Because we have a number of people working on this project (and hope
to have even more doing so in the future), it is best to work from
as recent a version of the database as possible when cataloguing.
This greatly reduces the chances that you will have to go back and
replace the new keywords that you have used while cataloguing.


DATABASE CHANGES

I went through the database and made some changes to make it
consistent with what we've said on the subject recently:

I changed "Analysis (posts dealing with how readers see and react to
canon)" and made it "Reader Reaction (posts dealing with how readers
see and react to canon)."

I deleted "Apparition" because we already had "Apparate."

I deleted "ENSLAVEMENT," "SOPE," "LOLLIPOPS" and "FLIRTIAC" because
they are acronyms.

I deleted "Get Fuzzy (comic strip)" because it sounds OT.

I deleted "Movie" because movie posts would be off-topic and should
be skipped for that reason.

I deleted "Troll (refers to the poster, not the creature)" because
it sounds OT. If someone later comes across a post discussing
trolls as magical creatures, feel free to add "troll" back in.

It might be best to go ahead and global replace now if you used any
of these deleted terms. If you feel that you are too far along to
make the corrections now (or if you're not comfortable with
the "Replace" feature), just start using the keywords in the
database from this point forward and we'll clean up your keywords
later.

Does all of this make sense and sound workable? Are there any other
general guidelines that we've missed? Let us know.