Hacker News Reader

Comment by onli | Hacker News Reader

It's kind of funny how specific some of those changes are. Issue #2373: Improvements around real estate description including the addition of a property accommodationFloorPlan and type FloorPlan for use when a number of accomodation entries share a common layout. Added numberOfFullBathrooms and yearBuilt. Okay then!

In general the schema markup is great though. It's semantic web stuff done right: It's not hard to implement, it has tangible benefits, and there is proper tooling around it like the schema markup tester. And best of all: While this is mainly about changing how search results look, there is nothing stopping other software from using that markup. It's really one of the positive things Google is involved in.

It's only frustrating if what you want to do is not supported. I'd love to markup my processor and graphics card benchmark results (the ordered result list), but there is nothing in the schema that would allow me to do that :/

View on HN

> It's really one of the positive things Google is involved in.

Is it though? Seems to me not necessarily a positive thing. If Google can serve the information on your website already on the SERP, the user won't have to visit your website anymore.

The trick is probably to expose enough to be useful and convince the user that you have more useful details on the full webpage. For example when I search something and get a full StackOverflow answer presented above the search results (both in ddg and google) that often leads to me following the link to read the comments and alternative answers to learn about the gotchas and nuance of the solution. Or to stay with the example of GP, you might tell google that a page is about an apartment or a hotel room with two bedrooms and two toilets, but anyone interested will still click through to see pictures and additional detail.

It also depends on the purpose of your website.

Are you publishing data to help/inform people? or do you want traffic on your site (for ads)?

Examples like Wikipedia is about sharing information, not so much about growing traffic..

> If Google can serve the information on your website already on the SERP, the user won't have to visit your website anymore

Isn't this the whole point of the semantic web - making information machine readable so that humans don't have to visit the web page anymore?

> If Google can serve the information on your website already on the SERP, the user won't have to visit your website anymore.

There are lots of services where getting the information to interested parties is the goal, and getting website visitors is only useful as a means to that go, rather than the other way around.

If that's not your interest, well, no one is making you use schema.org semantic markup.

You can write your own ontologies and host them, if you're so inclined. With enough `sameAs` links it will be a part of the integrated whole.

> I'd love to markup my processor and graphics card benchmark results (the ordered result list), but there is nothing in the schema that would allow me to do that :/

You're probably one of the most qualified persons in the world to say how such a markup has to look. Propose something at public-vocabs@w3.org or ask for it on https://github.com/schemaorg/schemaorg/issues?

(But of course once you do, Google - and others - could parse your data and deprive you of site visitors because they can answer queries natively)

>(But of course once you do, Google - and others - could parse your data and deprive you of site visitors because they can answer queries natively)

Isn't that basically the whole point? That information provided by sites can become independent from those sites? For scraping, re-rendering, modifying, etc

Right, but I think it's worth adding the disclaimer because there are also folks who are shocked (SHOCKED!) that something like this could happen.

But you're right, that's the whole point of schema.org.

I'm not sure that's true?

They have various values, like integers, etc all setup, They don't have a benchmark exactly, but they have Review, which can hold various values, such as:

PropertyValue or QualitativeValue or QuantitativeValue or StructuredValue

so you can record both opinion based values and discrete values about various properties. Plus Review holds an overall value(AggregateRating ).

Sure they don't have a discrete type "GraphicsCard", but they do have generalized things, with value based items you can store about a thing you are calling a Graphics Card.

So I think it's totally possible, but I've only spent a few minutes looking, so I may be missing something.

Oh, it's possible to attach the benchmark rating to a specific item. in my case I'm using https://schema.org/AggregateRating for that (since one meta benchmark result consists of several benchmark results). And the specific item is then just a Product, which is okay.

But how would you express that you have a list of those aggregate ratings, which are ordered on the same dimension? In general I'm struggling with how to create lists with the schema, and I'm not aware of them being used anyway. I might miss something.

But I think https://github.com/schemaorg/schemaorg/issues/2405 would cover it, a benchmark could be expressed as a ranked list type of a guide. It being an open and recent issue makes me think that probably it's really not possible to express yet.

Edit: I completely missed that this might already be merged, https://schema.org/Guide. I just missed it in the Changelog! It's part of the pending section.

I missed that as well. Glad you found it!

The amazing part of about json-ld is you can make up your own terms when you have missing terminology! And, if you care enough, you can push for them in Schema.org and get them introduced.

Reply on news.ycombinator.com