I've been using plucker on my Visor for about a month now and here
are my comments:
(I like the new design of the web-site; it is far better and
intuitive that when I was last there so there is no need to comment
on improving it).
* When I initially unpacked it and tried to use it I found it very
difficult to do what I was trying to do-- basically I wanted to
download a single webpage and load it into my Visor. I liked the
install procedure-- don't change that.
* Looking at the command names in bin/ I couldn't tell what the
commands did without running them; and when I did run them the error
messages were cryptic. (I still don't know what plucker-decode and
plucker-dump do and I don't care since plucker-build has all the
finctionality I need).
* The HTML manual (whether on-line or in pluckerdb format) is too long to
read there should be something in the README that says: "If you want
to download a page, do this...".
* I finally stumbled on the man-page which was great! I love
command-line stuff with millions of options. One nit I have is about
the inconsistent naming of the options: e.g. "--noimages" rather than
"--no-images". If "--no-urlinfo" has a dash after the "no", why
doesn't "--noimages"?
* --dbname should default to the contents of the <title></title> tag
of the document being downloaded.
* I wrote an zsh function as an alias for plucker-build, and I use
this consistently:
#pluck: download website to PDA
#takes 3 arguments: 1)URL to d/l 2)maxdepth 3)name of pdbfile
pluck()
{
if [ -z "$1" -o -z "$2" -o -z "$3" ]; then
echo "Usage: pluck URL maxdepth pdbfile "
else
plucker-build --pluckerdir=$HOME/public_html --home-url="$1" --maxdepth=$2 -f
$3
fi
}
I think it would be useful to have aliases pluck1, pluck2, pluckN
that would retrieve a doc, with maxdepth set appropriately to N
(based on stripping N from argv[0]).
* User-interface issue: I would like to suggest that links not
downloaded (followed) should either not be rendered in plucker or
should be rendered in another distinct manner that indicates to the
user, before he clicks on it, that it will not be followed. (If you
have a chance, read Alan Cooper's _About Face_ a good book on UI
issues.). It's just like a grayed-out menu option; the item is
there but it can't be selected rather than having a dialog box
pop-up saying that you shouldn't have clicked on this
link/menu-item, etc.). Perhaps if maxdepth could be embedded into
the document so plucker would know how many links deep it has data
for.
* Why do the bottom 2 lines of the previous page appear at the top of
the next page when scrolling? I've already read those 2 lines and
the screen real-estate is precious enough for this to be justified.
* Would it be possible to seamlessly read a large document rather
than ask the user to click "Click here for next part", etc. Isilo
does this and it would be nice if plucker this this too.
* There are 2 types of uses for plucker: 1) update a set of pages
"every morning" to be downloaded; 2) retrieve a specified document
on demand (something interesting on slash-dot, etc.). There should
be appropriate front-ends for these 2 distinct uses.
That's about it, for now :).
