@reiver@mastodon.social1/
I kind of get the feeling that people are trying to refactor the web, in a sense.
And that there are at least 2 different visions for the future of the Web.
Some seem to want the Web to go back to being about (relatively “dumb”) hypertext document format.
Others seem to want the Web to be about a (relatively “smart”) cross-platform application platform.
(There are other visions for the future of the Web too.)
2/
This difference between these 2 visions for the future of the Web reminds me of the “M” (i.e., “Model“ ”) and “V” (i.e., “View”) respectively from the “Model-View-Control” (“MVC”) software development pattern.
I.e.,
“Model” ⊃ “(relatively dump) hypertext document format”
“View” ⊃ “(relatively smart) cross-platform application platform”
3/
The ones that want the Web to become a (relatively “dump”) document platform seem to want something like the ‘Model’ from the Model-View-Control software development pattern.
The ones that want the Web to become a (relatively “smart”) cross-platform application platform seem to want something like the ‘View’ from the Model-View-Control software development pattern.
4/
I wonder if these two visions for the future of the Web have to be at-odds with each other —
The popularity of the Model-View-Control software development pattern suggests that they could be complimentary.
@reiver @atomicpoet I think that even on the web the two are complementary. The massive amount of effort that has gone into separating content from presentation, or the birth of µformats, is a testament to that.
Moreover, the losses we've had on some fronts (e.g. some resilience towards the adoption of some µformats by large proprietary platforms) always go hand-in-hand with attempts to destroy the web.
@reiver surely the fediverse is a good case study on this? At its heart, every member of the fediverse uses the same model (i.e. ActivityPub) but each implementation, and each client for each implementation, presents a subtly different view?
Isn't this basically what the API-first movement encourages? "Here's our view into our model, if you don't like it, grab the model yourself".
Of course, that's the complete opposite of what the Web 2.0 giants want with their lock-in culture.
@reiver Personally I have trouble with this analogy, but I guess it depends on what flavor of MVC you have in mind. I learned that “view” is supposed to be dumb, ideally logic-free, and “model” is where the business logic resides.
I prefer “document web” vs. “application web” but I agree they don't have to be at odds. I just think people treating the web as an application platform could save themselves a lot of pain by embracing web's hypertext nature.
@reiver The problem with the web is that it is a stateless protocol, making it inherently unsuitable as an application program. You need something stateful. In a way, that's easy ... just use telnet (or better ssh for security) ... like Bulletin Board Systems. Very little processing then needs to be done client-side. In a way, it's the perfect antidote to all the bloat. It does mean ASCII screens, though, and I can hear the outroar about that already.
Yeah, I agree that by itself HTTP isn't an "ideal" protocol for application development.
Although people have figured out some way of making-it-work to some degree.
Usually by communicating a "state ID" somehow.
(For example, via cookie, ETag, query string, special HTTP header, etc.)
But having things only communicated with a client-request & server-response model puts some limitations on it (i.e., HTTP as an application protocol) too.
But even with that, there have been some work-arounds to that, but they weren't ideal either.
(Long polling, persistent HTTP connections + JS, etc)
Regarding:
"just use telnet (or better ssh for security)"
It has been a while since I looked at the TELNET (TELNETS) protocol, but —
I think, at its core, it is pretty close to just a TCP connection, with a "special character" that is suppose to be escaped in a certain way.
Although there are some extensions to TELNET (and TELNETS) too, that create conventions for certain things.
I think if an application developer would divide up a TELNET (or TELNETS) stream into logical messages somehow, that websockets is somewhat similar to that.
Regarding:
"It does mean ASCII screens, though, and I can hear the outroar about that already."
In practice that probably, isn't an issue.
People just shove (Unicode) UTF-8 into anywhere that is suppose to be ASCII only.
And in practice it often tends to work.
@reiver I feel like a number of people want the web to go back to a time when it was simpler and less cluttered, but their view of that basically involves only text. I like Gemini and other smallnet protocols, but as an artist I've had trouble embracing them because I want to share images I've created but it's very difficult to do so. At least in a way that looks appealing and focuses on the images themselves and not the mode of presentation.
You are not alone with your complaint about the gemini-protocol.
The lack of built-in image support is a common complaint.
...
There are some other common complaints too.
...
I think there is an opportunity for a small-net format / protocol that addresses these common complaints.