1 |
Hi Sven, and all the interested people. |
2 |
|
3 |
I'd suggest to have possiblity of several solutions of the problem. |
4 |
Sometimes there is no good solution for the problem, but the number of dirty |
5 |
hacks with advantages and disadvantages in each. |
6 |
Another situation is when there is a problem with a certain version of |
7 |
software and there is two possible ways to solve it: |
8 |
1. upgrade to possibly unstable but newer version in which the problem |
9 |
doesn't exist |
10 |
2. Fix the problem |
11 |
In some cases 1st variant will be suitable, in others 2nd. |
12 |
|
13 |
On 5/19/06, Sven Vermeulen <swift@g.o> wrote: |
14 |
> |
15 |
> I believe that some interesting natural language engines exist already |
16 |
> which |
17 |
> might make a good candidate. One of the projects that I have investigated |
18 |
> thus far is the MindMeld project, which has a nice engine, giving good |
19 |
> results for both "how do I" questions as plain copy/paste of error |
20 |
> messages. |
21 |
> It actually builds the result based on both its database and prior queries |
22 |
> (with feedback from the user ala "was this answer helpful"). |
23 |
|
24 |
|
25 |
I've read about that MindMeld project. Its principle quite simple. |
26 |
I doubt that MindMeld's engine's data will be poisoned with common words |
27 |
specific to gentoo like "emerge","package" etc.They will be either cleared |
28 |
by statistical filter or overwiegh other keywords. |
29 |
|
30 |
From what I've read on MindMeld's site I've understood that they are using |
31 |
some kind of 1-layer perceptron model. This model is well-known and highely |
32 |
researched however in most cases it is too simple to consrtuct reasonable |
33 |
neural networks for human language analysis. |
34 |
|
35 |
And some IMHO about neural networks: |
36 |
I have to say that I'm quite familiar with theory and practice of wokring |
37 |
with neural networks.Their behaviour isn't very predictable. There is an old |
38 |
principle: whenever you can avoid using of neural networks, you should avoid |
39 |
it. |
40 |
I'd advice to look towards some other engines which doesn't deal with neural |
41 |
networks. |
42 |
|
43 |
Categories are quite difficult to use in a knowledge base. A more active |
44 |
> approach here is the use of keywords. If a topic has something to do with |
45 |
> the installation, the keyword "installation" is active. |
46 |
> |
47 |
> Of course, viewing all topics with a certain keyword can be seen as that |
48 |
> specific keyword categorie. All in the point of viewing. |
49 |
|
50 |
|
51 |
As I've mentioned in topic on gentoo forums, I'd like to see cathegorized |
52 |
keywords, possibly organised in some tree. Possibly there could exist 2 |
53 |
methods of searching: "by keyword" and "by natural language queries" or some |
54 |
mix of both of them (Possibly searching with "natural language query" and |
55 |
then filtering with keywords. |
56 |
Organising tree of keywords is not that hard as organizing topics itself. |
57 |
Even if you won't introduce full-featured keyword search it would be great |
58 |
to separate by keywords different types of problems, like "configuration", |
59 |
"install", "crash", "dependency" problems etc. This could be applied like an |
60 |
additional filter to the query results. |
61 |
|
62 |
|
63 |
The Gentoo Documentation Project should focus one other things, like general |
64 |
> guides, internationalisation of the available documentation, search |
65 |
> ability, |
66 |
> published articles, project documentation and the lot. |
67 |
|
68 |
|
69 |
Also one more thing came to my mind. Do you plan to allow |
70 |
internationalization for the KB project? If so, there is no point in |
71 |
sticking to some "natural language engines" that doesn't allow extension to |
72 |
the other languages. |
73 |
|
74 |
--------- |
75 |
Ты тонкий мыслитель сказал людоед, а больше толстых люблю (с) me |