Re: search-linuxkb progress report

[Aaron Turner <aturner@linuxkb.org>]

On Mon, 20 Dec 1999, Hugo Rabson wrote:

> I'm midway through flowcharting search-linuxkb. I would be coding it right
> now, had I not accidentally hosed my Linux installation. I don't have my
> RH6.1 CD at the moment & am having to download the RedHat tree via ISDN
> modem. (Remind me to buy a CD-writer.) I backed up /root and /home before
> the accidental hosing, so very little harm was done.
> 
> The commands available when calling search-linuxkb will be:-
> 
> # search-linuxkb --list (--simple|--advanced) (--wholepage|--programs)

I don't see any point in having a simple/advanced option.  simple/advanced
is a UI issue, not a content issue.  Advanced basically has more features
available to be configured by the end user, why simple just sets some sane
defaults.
 
> ...retrieve the simple/advanced search engine web page; write all / programs
> to stdout

Why are you retrieving the web page?  Aren't you building the UI yourself?

> ('programs' = the programs listed in a dropbox on the web page)
>
> # search-linuxkb --search (--simple|--advanced) \
> (--allmatch|--anymatch|--boolean) (--programs="...,...,...") <expression>
> 
> ...retrieve the search engine's web page; concoct the appropriate POST
> string (?a=b&c=d&...); call the web page with said string; format results;
> write to stdout
> 
> To keep unnecessary hits to a minimum, the script will keep a copy of the
> simple and advanced search pages. When a search page is requested, its copy
> will be used if less than an hour old; otherwise, it will be updated.

Again, it seems to me that the UI should be coded into the app.  I know
this may lead to breakage later down the road, but then again if you're
downloading the web page and then executing the same options on the site,
then what's the point?  Why not just start up netscape/lynx and access the
web site in that case?

I think we're loosing sight of the goal of this project.  At least, it's
definately changed from what I thought it would be.  I thought that this
was going to be either a library or command like application (stdin/out)
that allowed people developing other help applications to be able to send
queries and retrieve results from the KB.  Basically, it's not an end user
app, but a tool for other people to write apps with.

now the stdin/out can already be done with lynx:

lynx -(source|dump) http://www.linuxkb.org/search/remotesearch?cgiparams..

however lynx is pretty heavy (2MB) for something as simple as this.

I just don't see the point of writing an end user application that only
queries the KB, that already exists as the web browser.

Comments?

--
Aaron Turner, Core Developer       http://vodka.linuxkb.org/~aturner/
Linux Knowledge Base Organization  http://linuxkb.org/
Because world domination requires quality open documentation.
aka: aturner@vicinity.com, aturner@pobox.com, ion_beam_head@ashtech.net