Keyword Questions, Protocol Guidelines, and SHIPping

Sun Oct 6 02:25:49 UTC 2002

Elkins claws her way out from under a tottering pile of manilla 
folders, each one labelled with words like "Boggart," "Train," 
"Potion" and "Name," and staggers to her feet.  Halfway across her 
office, she trips over the stuffed-to-bursting "Ever So Evil" file, 
stumbles, pinwheels wildly for a few moments, ricochets off of the 
large trash bin labelled "Plot Hole That Must Not Be Named," and 
finally comes to a rest lying half on top of the paper shredder 
marked "Hey,what's With That Briefcase, Anyway?"  She closes her eyes 
and moans.

It suddenly occurs to Elkins that it might be nice to just die here.  
They could find her body like this, draped over the Briefcase File, a 
scrap of paper marked "Does anyone know what the J. stands for?" 
still clutched in her cold stiff fingers.  

But no.  No, that just wouldn't do.  There is work to be done, after 
all, and besides, there have probably been a couple of posts to the 
FAQ list in the past week.  Just a couple.  One or two. It's even 
possible that people have actually begun cataloguing -- although 
really, that's hardly likely, is it?  No, probably no sense in being 
too optimistic here.  Optimism only leads to disappointment.  

Elkins sighs.  She pulls herself wearily to her feet, shuffles across 
the room, pauses to shift the "Non-Compliance" file out of the way, 
and yanks open the door.  

At first the noise and bustle and light is all just a bit too much.  
After a week of retreat, it registers as little more than random 
sensory input.  Only gradually does Elkins begin to make sense of 
what she is seeing.  She stares. Her jaw drops.  Then she falls to 
her knees and bursts into tears of helpless gratitude and relief.

--------------------------

Wow!

You guys are truly, sincerely awesome.  You have been doing such 
great work, all of you.  I don't even know how to begin to register 
my excitement over how this project is going.

It's a bit late for me to welcome Porphyria, Eloise, Pip, Debbie and 
Eileen, but I'm going to do it anyway.  <waves to killer catalogue 
team>  

Gwen and Porphyia, great job with SnapeFAQ!  I've just read over the 
docs so far, and they look fantastic.  Cindy, the History FAQ looks 
great.  

Dicey, we Luddites would be lost without you!  I know that I've been 
immensely relieved to see many of the so-called "silly questions" 
that people have been asking here, because I had the exact same ones 
myself, and now I'm spared the embarrassment of having to ask them.  
Your explanations have been super-helpful, and the "how to use Excel" 
document is a work of beauty.  Thanks so much.

I hope that Eloise's children are better, that Pip's computer has 
shaped up, and that Cindy and Porphyria enjoyed their anniversaries.  
My own wedding anniversary, which was on October 1, very nearly 
passed me by but fortunately, this year I had HPfGU to remind me of 
it before the day had ended.  Even more fortunately, I remembered it 
*before* the Spouse did this year -- which is a first for me.  Then, 
we're neither of us too good at that sort of thing.  We usually don't 
remember it until sometime around the middle of October.  Still,
though, the Spouse usually gets there first, so I was pleased to beat 
him out on it this time around.  ;-)

On the LupiFAQ front, the preliminary culling of the archives is now 
completed.  Amy and I should be moving on to the arranging stage next 
(I've sent you all of the Briefcase posts to sort through, Amy.  Hope 
you don't mind. <eg>)

I've also starting my own cataloging today, by the way.  The list was 
very obliging in reaching message #45,000 on *precisely* the same day 
that I returned.  It might have been nice to have a day off, of 
course.  Oh, well.  Never mind.

It's taken me some time to finish playing catch-up on everything 
that's happened here.  Now that I'm caught up, here are some thoughts 
on a few of our more pressing protocol issues.  Nitpicks and trivia 
to follow.

---------------------------

JUDGEMENT CALLS

First off, just a word about judgement calls (or "judgment calls," 
for all you lawyers out there).

There's been a lot of back and forth about certain keyword issues 
these past few days.  That's not a bad thing at *all.*  It's great 
that everyone is feeling so involved and enthusiastic about the 
project, and a lot of the exchanges have been very useful in terms of 
making sure that we're all on the same page, especially when it comes 
to reaching a shared understanding of how we're using terms 
like "redemption," "prediction," "prophecy," and so forth.

I would like to suggest, though, that we might want to try not to get 
*too* stressed about all of the consistency issues at this stage in 
the game. Obviously we don't want to be all over the map in our 
keywords, but if there's a bit of variation, it's really not a 
disaster.  Once the catalogue is completed, we're all going to have 
to sit down with the keyword list and discuss consistency issues 
anyway, so if there are minor variations at this stage of the 
project, it's not a big problem.  We will have the opportunity 
to iron out all the wrinkles later on.

Also bear in mind that to some extent, responsibility for actually 
*using* this catalogue is going to have to fall on the shoulders of 
the FAQ editors themselves.  We do want to make it as easy as 
possible for the FAQ writers to find relevant posts, yes.  But the 
FAQ editors themselves will also need to have the presence of mind to 
use the database sensibly in order to find what they want.  

If a message about the judicial branch of the MoM, for example, ends 
up being marked with keywords "Law," "Crouch," and "Auror," then it's 
really not a disaster if it is not also marked with the 
keyword "MoM."  Honest.  It really isn't.  It's fine, because any 
sane person looking for posts about the MoM's judicial system would 
think to try the word "Law."  The end users are not going to be 
looking for just one keyword and leaving it at that, so there's no 
need to get too worked up about making sure that every conceivable 
keyword that could possibly be applied to a given post gets entered.

In short, while we naturally want to be both consistent and 
descriptive, we can safely assume that the people who will actually 
be using this document are not going to be *brain-dead.*  

So I don't think that it's necessary for us to get too flustered over 
minor variations in keyword designation.  If there's a bit of 
variation, that's okay.  We can fix it later.  Let's try to focus our 
attention on Big Fat Problematic Inconsistencies for now, and save 
the tinkering for later. 

THE KEYWORD DATABASE 

This is a great tool, and it makes for fun reading, too.  I've just 
started my own chunk today, though, and I've found that having a 
plain old text list of the keywords to date printed out and sitting 
beside my computer has been very useful.

Unfortunately, if there's a quick and easy way to know when something 
new has been added to the database, I'm unfamiliar with it.

So just a request.  If people could post to the group itself whenever 
they make any new additions or changes to the database, that would be 
very helpful.

I know that you've all pretty much been doing this anyway, but I 
don't think that it was ever stated explicitly (unless I missed it 
somewhere in the hundred or so messages that I've been catching up 
with).  So I thought that I'd make it a more formal request.  It will 
be *much* easier for me, and probably others, to keep up to date on 
the current keywords if people's changes or additions to the 
database get posted to the list as well.

Debbie asked:

> Topic terms: Do we want new topic terms to be posted here for 
> debate/approval? Or should we just upload them to the list if those 
> subjects aren't yet covered?

Cindy replied:

> I think we should just upload them, and only raise questions when 
> we aren't sure the best way to go about things.  

I agree.  I'm not sure if I think that too much quibbling over 
keywords is really all that productive a use of our time.  We're all 
smart people with good judgement, we're all here by invitation, and 
we all trust each other (I hope).  So I think that unless we feel a 
strong need for a second opinion, we should all feel free to upload 
new terms to the database, and then post to the list to inform 
everyone of what we have done.

Practicing what I preach here, I have just added the following terms 
to the database:

Book Five (OoP)
Theory (literary criticism)

I have also edited for spelling:

"Plagarism" has now been changed to "Plagiarism."

Eloise asked:

> Are we supposed to list *every* topic term we use?

Only ones that are not mind-numbingly obvious, I'd say.  In other 
words, I don't think that we need database entries for the names of 
characters.  We all know that "Harry" is a keyword.  Same for "Snape."

Anything less obvious, though, then I'd say go right ahead and add 
it.  It's probably better to err on the side of caution here.

GRANULARITY ISSUES

A good deal of the discussion here seems to have centered on 
granularity issues: how specific should keywords be?  

I think that here it may be useful to try to keep in mind how this 
database is likely to be used, and for what purposes.

To give an example of how this principle might be applied in 
practice, let's take the question of relationships between 
characters.  How should they be designated?

Well, okay.  Let's say that I was writing the Dumbledore FAQ, and I 
wanted to find a good post on Dumbledore's relationship to Snape.

The first thing that I would do would be to ask Excel to sort by 
"Dumbledore."  This would turn up all of the entries that used 
"Dumbledore" as one of the keywords.  (I would probably already have 
done this, actually, but soft. This is a hypothetical.)

The next thing that I would do would be to tell Excel to sort the 
next columns for "Snape."  This would pull to the top every 
"Dumbledore" entry that also had a "Snape" keyword.

Then I would read the "comment" fields for those posts to see which of
them contained the sort of subject matter that I was looking for, 
keeping an eye out for descriptions like "Snape's relationship with 
Dumbledore," "Dumbledore as Snape's father figure," "Was Snape told 
about Sirius as animagus?" and so forth.

Now, if (for some sick and twisted reason) I was specifically looking 
for posts discussing the possibility that Snape and Dumbledore 
are ::shudder:: *lovers,* then I would narrow my search down still 
further, by asking Excel to sort not only by "Snape" 
and "Dumbledore," but also by keyword "SHIP."  

So neither Dumbledore-Snape nor Dumbledore/Snape is necessary as a 
keyword.  In fact, their use could be damaging, because it would 
prevent those posts from turning up on a simple "Dumbledore" search.  
If somebody wants a list of every message dealing with Dumbledore, 
then we want them to be able to find them all by simply sorting on 
keyword "Dumbledore."  Keyword "Dumbledore" is not the same thing as 
keyword "Dumbledore/Snape," and keyword "Snape/Dumbledore" is worse 
still (from the point of view of the Dumbledore FAQ editor), because 
it wouldn't even be anywhere *near* to "Dumbledore" in the 
alphabetical array.

So we really don't want to get too specific with the keywords.  When 
in doubt, it's always better to add on more keywords than to try to 
make the keywords themselves more specific.  The place for 
specificity is in the "comments" field, where the more full a 
description you can give of the argument of the message (within 
reason), the better.  The keywords themselves, however, should be
reasonably broad.

Maybe a bit of brushing up on what spreadsheet programs are capable 
of might be in order here?  I'm a Luddite myself, but after I 
finished churning my butter this morning, I played around with Excel 
a bit, and it can do a *lot.*  I think that maybe people aren't quite 
appreciating the sort of sorting functions that the program itself 
can take care of for us.

Amy, for example, asked:

> My big question is how specific to get in creating categories. . 
> . . First post, topic: Ginny.  It's a general post (not very 
> interesting, truth be told) and I thought, what if I were using 
> this catalogue to write a FAQ on Ginny?  Wouldn't it be great if 
> the catalogue did what I was going to have to do otherwise: not only
> find all the Ginny posts but put them into categories?

As I understand it, that's what the spreadsheet program is for.

Excel can do that sorting for us.  That was the reason that in the 
end we decided to go with Excel for the final product, rather than 
with Word or any other word processor.

> Or--<holds breath>--will there be a way for users of this catalogue 
> to search for cross-listings, so that anyone looking for posts on 
> Snape's loyalty can seek out those posts that deal with both Snape 
> AND loyalty and scoop 'em all up like that?  

Yup!  Sure is.  So you can let your breath out now, Amy.  ;-)

Yes, that's how it works.  If you wanted to look for posts on Snape's 
loyalty, you would tell Excel to sort by "Snape," and then by 
"Loyalty."  That's the program's function.  

It's beautiful, isn't it?

So, yeah.  Don't try to get too specific with the keywords 
themselves.  As counter-intuitive as this may seem, the final product 
really won't be nearly as useful if the keywords aren't relatively 
broad in scope.

Now, as for how this relates to shipping...

Cindy and I have discussed this issue, and we've come to the 
following decision.

SHIPPING POSTS

Eloise suggested:

> How about we take a ruling from higher authority on that one?

And Eileen agreed, crying:

> We need a ruling on high about SHIPs! Complete with guidelines,
> examples, etc. 

Okay.  You want a ruling from on high?  

Here's a ruling from on high.

<Elkins stands on her tippy-toes in a rather sad attempt to look like 
Somebody Important>

Relationships between characters should not be expressed by means of 
the keywords at *all.*  Leave that to the content field.

A LOLLIPOPS post, for example, might have keywords "Snape," 
"Lily," "James," "SHIP," and (if it were TBAYish) "LOLLIPOPS."  

The description of the post could then read: "Brilliant LOLLIPOPS 
defense," or "Snape loved Lily," or "Snape not only loved Lily, but 
she loved him back; the two of them were having assignations in the 
broom closet up until their very last year at Hogwarts," or 
"Snape is a half-dementor *and* he loved Lily," or whatever.  But 
leave the keywords generic.  No slashes.  No hyphens.  None of that.

I know that this contradicts what we decided earlier, and if you've
already entered a bunch of H/H keywords, please don't worry about 
it.  We can fix them later.  But from here on in, please don't use 
slashed keywords.  We already have a SHIP keyword to help people to 
find the shipping posts, and slashed-or-hyphenated keywords really do 
cause a lot more problems than they solve.

Eileen complained:

> I am going to go nuts having to write Keyword 1: SHIP Keyword 2: 
> Ron Keyword 3: Harry Keyword 4: Hermione, Keyword 5: Ginny, Keyword 
> 6: OBHWF, Keyword 7etc: whatever the post addresses in relation to 
> OBHWF for simple posts about OBHWF.

Yeah, I know.  It's a real pain, isn't it?  Sorry, but it can't be 
helped.  

Eloise asked:

> Should I put in a SHIP key when the writer for example, refers to 
> LOLLIPOPS but to refute it? Negative shipping as it were? I haven't 
> done, but I suppose it gives the other side of the argument that 
> someone writing about a particular SHIP might want to include.

Yes.  Posts refuting a SHIP argument should still get a SHIP key, just
as they would get a SHIP prefix on the list.

Again, I know that this makes for more work, and I'm sorry.

Eloise:

> I've found quite a few of those sort of posts both deal with an 
> individual *and* with that individual's relationship with someone 
> else, sometimes in some complexity. Maybe it's not necessary to 
> specify that, but I just thought that if, say writing a FAQ on 
> Sirius, it might be quite helpful to be able to identify easily 
> those posts that explored the naure of his relationship with 
> Snape or Lupin, for example.

Specify it in the content cell.  That's what the content cell is 
for.  

This catalogue should make the FAQ writers' jobs *much* easier, but 
it's still not going to do all the work for them.  A keyword search 
for both "Snape" and "Sirius" should turn up all of the messages 
dealing with both of these characters.  Adding "Prank" to the mix 
would make the search more specific still.  Beyond that, though, the 
FAQ editor is on her own.  She can glance down the descriptions to 
find messages that match any more specific content requirements that 
she might have.

Eileen:

> This is beginning to remind me of the time when I was 14 that I 
> decided to catalogue my parents' collection of battered National 
> Geographics since the 50s. I had a very complicated system for 
> that.  It wasn't the most successful of endeavours.

<Elkins winces in memory of her own, very similar, adolescent 
experience with cataloging, then shakes her head firmly>

Don't worry.  So long as we follow the KISS rule, this one will work 
better.  The key here is simple keywords.  Very simple keywords.  
Don't try to do the FAQ writers' job for them.  The catalogue's 
function should be to serve as a useful tool for FAQ compilers.  We 
should not be trying to compile the FAQs ourselves.

Not yet, at any rate.  That comes later.  ;-)

Elkins (who as a student found her University work-study job as a 
department secretary so mind-numbingly boring that she decided to 
catalogue the department's entire library for them instead, and who 
swore afterwards that she would never do something like that for 
minimum wage ever again.  O, the irony.)