Dave Winer recently wrote:
Twitter is the new OS. For the moment. It’s like the Apple /// in the days before the IBM PC. Remember the Apple ///? (Not many do.)
I think that’s a good insight, and I like the comparison to the early days of the PC (really the pre-PC). But I don’t think Twitter is it.
What I think is coming is going to be more like a jumble of Greasemonkey and Shiftspace. Though I don’t think either of those are “it” — at least, not yet.
Those services still look, to me, like that Altair 8800. Greasemonkey in particular. It’s a platform, or a set of tools, or something that you can use to do stuff, but it really doesn’t do anything on its own. Started and still primarily used by hobbyists, uber-geeks, web-nerds.
Cringely began his “Triumph of the Nerds” mini-history of the PC with an episode called “Impressing Their Friends”, basically arguing the thesis that we owe the phenomenal technology we have today (mid-nineties, when the series aired) all because a group of hardcore nerds spent a lot of time trying to impress their friends.
Compare that to this from this 2005 Wired article on Greasemonkey (probably the only mainstream magazine article about Greasemonkey so far) (emphasis added):
Greasemonkey was originally written by Aaron Boodman, who wrote the program in December 2004 to amuse his friends and found himself pleasantly surprised when it grew into a cult hit.
I think 10 years from now, when people talk about how they are using computing, Greasemonkey is going to be singled out as a sort of turning point, like the Altair 8800. Greasemonkey itself will probably not ever be used by the millions and millions of people who surf the web (neither did the Altair 8800 enter into “mainstream” usage); I imagine it will remain the playground of web-hackers and early adopters.
But I do see a prominent place for something building on what Shiftspace has created; a “layer” over the web. It will be easy to use, easy to install. It will probably be standardized in some way. It will allow people to see layers of annotation, comments, user-created links, and additional content over the entire web.
The owners of any given website will lose control over what their visitors see or do not see; a visitor to your site might choose to view it with only the content and styles you provide. They may choose to see what annotations & comments (if any) a certain group has been making on your page. They may choose to highlight a phrase from your blog post and directly annotate with a long essay on why you’re wrong (or right). They might click on the user name of another annotator and be taken to a page that lists all the recent annotations and interests of that person.
The people on the bleeding edge of what’s new and hip on the web don’t seem to be very excited about the scenario I’ve just described. So far, I’m thinking there are two possible reasons for this:
- I’m wrong. This will never happen, and no one cares. Obviously, I don’t think this is the right reason, but I’ve been wrong before, so I guess we can’t rule it out. Or,
- It’s not the Next Big Thing. It’s the Next.times(n) Big Thing, where n > 1. I don’t think of my self as being especially prescient, so I think if this is the case, n is still probably a pretty low number. 2 or 3 at the most.
I like Shiftspace, but I’m not sure it’s the application of this concept that is going to break through to the mainstream; though I think it could be the application that opens peoples’ eyes to what is possible.
Something like the Apple II did for the personal computer.
I dunno. FWIW (not much, I know), I can't see greasemonkey ever reaching wide acceptance unless the owners of the site decide to play along with it. Which, if they were going to do that, they'd just wrap the stuff up into their own site anyway.
I've only really used greasemonkey a few times. Every time I used it I wound up being frustrated within the week – the script relied on some of the site's internals that changed as part of an upgrade.
Applications – especially web applications – are moving targets when you're working outside of a contract (e.g. API). That makes anything outside of a contracted change to the site only worthwhile until the next change happens.
I imagine the same might apply to Shiftspace, in some degree, but I haven't ever used it so I can't comment on that one directly.
Well… that's actually what I said, sort of. I don't see Greasemonkey hitting the mainstream either.
What I'm thinking about is something like I describe a few paragraphs before the end; a sort of shared layer overtop the web. Shiftspace already has a kernel of something like this operational. It's absolutely possible. It will only get more possible.
The fact that people are not jumping off the walls over this idea doesn't persuade me that it won't happen… only that it will be the Next Next (Next) Big Thing… rather than just the immediate “Next”.
Regarding, moving targets. This is true, and is a major challenge. However, as this becomes more mainstream (it will sooner or later… the ability to comment/annotate anything, anywhere, is just too compelling, and completely feasible), it will eventually become incumbent on the regular web that it is better to standardize and allow “hooks” that don't change. Much like an API, of sorts.
In case this sounds too far-fetched, keep in mind that Gmail has already done this; See Gmail Greasemonkey API. To quote them:
Again, I don't think it's a question of “if” this goes mainstream. It's only “when”. Maybe it's a couple iterations of Next Big Thing's away… but that's not really that far.
Yeah, forgive me my brash commenting. It's been a long day ;)
I only meant to offer my experience with greasemonkey as evidence of the moving target problem.
re: Google… what happens when people want to change things that aren't provided by the API? When Google makes changes to the API at some point, are they going to work with the greasemonkey devs to make sure their scripts are up to date when the new version launches?
Isn't Google's offering really lop-sided since it leaves out the 60% of users that don't use Firefox?
IMHO:
Greasemonkey is powerful because it gives you total control. Once you limit that control to a predetermined set of features – an API – you're basically just including a fancy configuration engine.
I think Google made the odd – as opposed to inevitable – choice of exposing those configurations to the user as part of a scripting engine. Why is it odd? You don't have any control over it. In making it official you've made the greasemonkey scripts something that people might expect support for. Just like developing javascript for a webapp – at some point, someone is going to install some other script that interferes with yours, and try to convince you that yours is broken.
It's a headache. I'd wager that most places are going to add in configuration for what they feel should be configurable, and then practice laissez-faire for the rest.
I've always thought it a bit puzzling that Google quoted the borg when they announced that API. If resistance is futile… Why are you putting in measures that attempt to define what and how people do things to your site?
I can see all your points, I think. To me, all it points to is that what I'm picturing is perhaps more than several iterations away from becoming mainstream.
And again… it's not Greasemonkey. Greasemonkey is to my imaginary web layer what the Altair 8800 is to the modern PC or Mac… not even a kernel, just a hint of what could be done.
Or maybe I'm just daft. But, whichever, it's fun to think about. :-)
Nah, I could see it happen. I just don't think it'll be particularly easy, and it'll probably be fragmented at best for a very long time. Nobody likes to make and/or adhere to standards anymore, it seems ;)
And while I get that you're not talking about greasemonkey… that's the example we have, and there are certain abstractions I think you can take away from that (specifically, what people are trying to do with it). That's all I'm after.
Shiftspace is a much better example of what I'm thinking. It's in version 0.10 and runs only on Firefox 2.x, but if you have that browser around, and this sort of thing interests you, it's worth looking at. Installs in about 3 clicks. I haven't used it in a long time, since I started using the Firefox 3 betas, but I'd highly recommend it.
Thanks for introducing me to ShiftSpace. I hadn't ever heard of it and it's been making me ponder all afternoon.
I really like the idea of having your own “view” of web pages… remolding them to fit your particular preferences when you apply your view, or using the page as a datasource to create your own visualization of what information it contains, possibly linking it with other “sources” in other pages and creating a semantic web in that way.
You could even have views intended for particular people. Say, I had a view I created of my blog for my family to see and another for the people I work with. Not that they'd have to use to the views I offered, but maybe they would.
Comments, highlights, annotations, links, context – all those things could be kept in those views, hidden or shown on demand.
And yes, there would be complications as a good bit of the web pages on the internet are dynamic, so in that case there would either need to be a really good context-based dynamic mechanism for interacting with the page or something like an API established as you and Duwanis were discussing.
I agree it's probably not the Next Thing as you say and probably a few more Nexts from the present but I can definitely see it as a possible future evolution for the web.