UPDATE , Sunday, September 5th

carolynwhite2 carolynwhite2 at aol.com
Sun Sep 5 12:07:01 UTC 2004


PROGRESS
As mentioned last week, Paul has created a new analysis tool to help 
me keep track of what is happening on the database. Amongst many 
useful things, it allows us to see the net numbers of posts rejected 
and coded up, without the element of double-counting that was 
occurring previously due to:
(a) posts rejected but also getting a code in the admin section 5; 
(b) posts included in numbers of coded, but in fact ignored as they 
appeared 'corrupted' on your screens. 
(In fact, Paul now tells me that this type of text indicates that a 
post has been removed from the original archive by its author).

So, using the new system, this week we have coded/allocated for 
coding 25565 posts, representing the 7815 posts in the old Yahoo 
club, and up to post 17750 on the main list.

Out of the 25565 posts, we have actually coded 23768 posts, and out 
of the 23768, we have rejected 15270 - 64%. (It is noticeable how the 
reject rate is now beginning to fall steadily).

This week, with 6 people coding, we managed 771 posts.

NEW ANALYSIS TOOL
Those of you who are idly curious about statistics might like to look 
at the analysis screen for yourselves. Unfortunately, the numbers you 
see are never likely to exactly correspond with the ones I give you, 
as they change as soon as anyone does any work on the database 
(obviously), and I am just grabbing them at a single point in time, 
but you might find it interesting anyway.

To go and look at it, first get into the catalogue, and you will see 
that the catalogue address looks like this on your browser:

http://24.0.253.65:8888/admin/categorize.php

Now replace the word 'categorize' with the word 'stats', being 
careful not to delete any other part of the address (slash, full stop 
etc). Highlight the whole amended address and hit GO.

A screen should come up showing you the current number of posts coded 
to each category, and various totals at the top. Please note that 
these totals are *not* the totals of the three columns to the right 
of the screen - they are adjusted to remove duplication, as explained 
above.

Also note that within this screen you can hit 's' against any 
category and it takes you to the screen showing exactly which posts 
are making up the numeric total you were looking at. Closing this 's' 
summary keeps you within the stats analysis.

To get out of the stats analysis altogether, you can either close it 
top right, which will crash you out of the catalogue and you will 
have to re-enter as normal. Or you could replace the word 'stats' 
with the word 'categorize' [Brits, note the spelling!], highlight and 
hit GO to take you back.

One thing I am finding this very useful for apart from the overall 
stats is noting where mistakes in coding have been made. For 
instance, in column 2, there should only be zeroes between reject 
section 1 and admin section 5. Where I found a number instead, I have 
been able to identify the wrongly-coded post and go in and correct it 
[I will post a full analysis of what I have done so you can go in and 
check your own posts if you like].

Similarly, there should only be zeroes in column 3 against the reject 
codes. Right now, there are quite a few numbers and I am working on 
correcting these today. I will include them in the analysis I post 
summarising what I have done.

There are other things we can use it for too - for instance to check 
that posts have not been coded to headers, where we have agreed we 
will only use the sub-heads. Eg under 'predictions' - we must split 
them between canon/non-canon, not bung them into the header box.

I guess that is much more than you wanted to know, and undoubtedly 
qualifies me for pending LOON membership.

CATALOGUE GROUP
Speaking of membership, some of you may have noticed that our group 
has gone down to 19 from last week's 20. This is because Jayne (The 
Sergeant Majorette) has decided to spend more time with her remaining 
sanity, and has bowed out. She has nevertheless offered to do any 
admin work we can think of on the catalogue, 'as long as it doesn't 
involve reading posts' (!). As she really does know how to use a gun, 
who am I to argue...but I have thanked her warmly for her 
contribution so far. I'll be looking for a replacement over the next 
few weeks.

NEW CATEGORIES THIS WEEK

1.1.3.3 (1097) Forgiveness
1.4.6 (1093) Family dynamics
1.4.6.1 (1094) Weasleys
1.4.6.2 (1095) Blacks
1.4.6.3 (1096) Dursleys

2.12.16 (1091) Spiders
2.12.16.1 (1092) Aragog

3.8.3.1 (1100) Wand-less magic
3.16.1.2 (1099) Room of Requirement
3.16.12 (1098) Hogwarts Terms & Holidays

And finally, on the subject of categories, please could you respond 
to my query earlier in the week about sins and virtues ? Its not the 
first time I have asked the question, but the subject keeps coming 
up, and might mean quite a lot of sorting out of the meta section.

Thanks

Carolyn






More information about the HPFGU-Catalogue archive