[Greasemonkey] a difference in vision

[Jeremy Dunck]

On 5/27/05, Matthew Gertner <matthew at allpeers.com> wrote:
> 1) ...in an ideal world, the
> capabilities that GM provides would be inherent in the Mozilla extension
> mechanism, which would have the advantages of Greasemonkey (rapid
> development and deployment, ...
...
> What do I think as of right now? We should do 1). It would require
> coordination with the Mozilla team. Does anyone at all agree with this? Even
> a little?

I'm all for lowering barriers to extension dev.  The 1.1 no-restart
thing sounds fantastic.  You still have the problem of really, really
fat topics, a dearth of documentation, and a relatively small niche
community (as opposed to page DOM dev pool).  (Side note, my next
requirement after no-restart overlays is a chrome reload button.  The
dev cycle on user scripts is as tight as it is because I can totally
reload the execution context easily.)

I don't think extension dev will ever replace the need for GM, because
the context of a user script (JS + page DOM) is smaller and more
familiar to many, many more people.  Unless there becomes a social
apparatus to focus down on "page extensions", or however you want to
put it.

There are a lof of folks that want GM to be everything extensions are,
without the headaches of extensions.  Obviously, making extensions
both easier to develop and more approachable would allow GM to be good
at the smaller problem of knock-off scripts shared freely.

An ambivalent goal for me is to allow users of other browses (IE?) to
do user scripts, and I think it'd be a shame if the mindshare went to
user script=firefox.  But that's more political than technical.  And
besides, I want Firefox to win, so why do something cool in IE?  Heh.

> 2) The whole principle of GM is way too fragile for what people want to do
> with it, but this has *nothing* to do with GM. 
....
> Regarding 2), there is going to be an evolution over to XML that will make
> the world a happier, sunnier place. 
...
>I think that a structured approach to screenscraping HTML ->
> XML combined with XML GM will be a big step in the right direction.

I think it's useful to separate data source code from data consumer
code, at least on popular sites, but I am highly skeptical of there
every being the grand stable contract you describe.   There's just not
enough of a incentive for the folks with the data.

An obvious exception is governmental offerings, in which we can pay
for however much stability and control as we like.  OK, I'm smoking
crack... ;-)