Comment by Lerc | Hacker News Reader

Comment by Lerc | Hacker News Reader

I think the fundamental problem with semantic web content is that it is binding two things together.

1. This feature does X

2. This feature is for Y

Misuse happens when people want the functionality for another purpose. I have seen some attitudes that are strongly against doing this even going back to the days of WinAmp where people were arguing passionately that custom user interfaces like that should not be permitted. On the other hand lots (I would guess millions) Opted for WinAmp precisely because of it's UI customization. It wasn't for everybody, but optional things do not need to be for everybody.

Binding what an element does to what it is for pollutes the semantic space whenever someone wants the functionality for another purpose. They end up with a feature that does one thing and broadcasts that it is supposed to be doing something else.

Perhaps an attribute could assist in this area (perhaps offlabel or something similar or more appropriate)

<meter min="0" max="250" value="180" offlabel>???</meter> To indicate that that an element performs functionally like a meter, without any semantic assumptions.

Given the greater degree of awareness that a developer should have when developing accessible content. I think it would be better to have the default to be the raw functionality and an attribute to indicate that the semantic context is valid. I don't believe that proponents of the semantic web would go for that, because not enough people would opt-in. They would prefer opt-out or no option at all.

The core concept of semantic web is that a “feature” _doesnt_ correspond to a behavior.

You can’t use emphasis to show italics, because it is used for emphasis- not a specific type of emphasis. And you as the author of a semantic HTML document don’t know how emphasis will work on the final users rendering.

I just wanted to point out that both you and the one you're replying to keep writing "semantic web", but I think what you really mean is "semantic HTML".

The semantic web is the umbrella term for the stack of technologies that include RDF, SPARQL, OWL, etc. — it doesn't really have anything to do with HTML, other than some special syntax for embedding RDF into HTML.

You are right that parent and grandparent should have used semantic HTML instead of web.

But I’m not sure I agree on your definition either. Looking at Wikipedia, semantic web is described, first as Web 3.0 (not to be confused with web3) but I remember it more as a dream of late 90s early 2000, where all data lived on internet accessible urls and had semantic markup, than a tech stack.

Well, it's a pretty concrete tech stack these days even if the original vision didn't take off.

I actually use these technologies in my day job (and have been for several years), so it's not just a wikipedia look-up for me. Maybe that's why I think of it in more concrete terms.

>Web 3.0 (not to be confused with web3)

I find this distinction to be rather ridiculous. They both semantically(heh) proclaim to be the next version of the web. I see there's a bit of an edit war on https://en.wikipedia.org/wiki/Semantic_Web over the one true web 3.

It's just an arbitrary name where people have incremented a number for something they wanted to be the next big thing. Realistically I think you could only really numerically name the next thing after it has actually successfully been adopted.

I had pretty much forgotten most of those things had existed. I see someone has sneakily added ActivityPub to the end of the list on wikipedia in 2021, retroactively giving them a success.

I assume there are pockets of existence for these things but for the most part the only time the majority of people encounter anything under the semantic web umbrella is where it has influenced html. It's a tomato is a fruit or vegetable thing, it is technically fruit but to most people it is effectively a vegetable.

Semantic HTML and the Semantic Web are not really related, other than both of them having "semantic" in the name.

> I see someone has sneakily added ActivityPub to the end of the list on wikipedia in 2021, retroactively giving them a success.

Why retroactively...?

The point of these standards is that they all build on top of each other. ActivityPub builds on JSON-LD which is a variant of RDF which uses the existing RDF schemas to define its attributes.

The different bits of the stack are used in many different ways. RDF datasets are really common in many different academic fields and ontology composition (which has always been a thing) has pretty much been subsumed by the semweb standards.

Semweb standards are also the most common standard in graph databases.

New W3C standards, such as ActivityPub, appear all the time and usually build on the existing semweb standards.

This is the is/ought distinction again, isn't it.

> I have seen some attitudes that are strongly against doing this even going back to the days of WinAmp where people were arguing passionately that custom user interfaces like that should not be permitted. On the other hand lots (I would guess millions) Opted for WinAmp precisely because of it's UI customization. It wasn't for everybody, but optional things do not need to be for everybody.

This really needs to be printed out and displayed on a billboard.

The existence of options, and other people making choices, is not a threat. There needs to be space for nonfunctional whimsicality. A music player is a good place for it. An aircraft console is probably not a good place for it.

Reply on news.ycombinator.com