Hacker News Reader

Comment by PommeDeTerre | Hacker News Reader

Why do we need two different syntaxes for representing trees? Both XML and their so-called "HSS" are redundant. S-expressions are a much cleaner syntax, and can be used to represent both the "markup" and the "styling" in this case.

For crying out loud, let's finally unify the syntaxes, rather than creating yet another way to represent tree structures.

View on HN

Just to make it clear, HSS is directly replacing CSS. No markup there.

You are missing the point, unfortuantely, the whole idea of the project is to sepparate the data from the presentation. Mixing them together would be pretty similar to what HTML was in 2005. Not desirable.

I believe you're missing his point: he said unify the syntax, not mix the data with the presentation.

Oh, I totally understand. Thanks CodeMage. Sorry for the misunderstanding.

XML is a widely-used data interchange format. It is well known and used. It's focus is the actual content.

The syntax of HSS/CSS focuses on attributes. Makes for a perfect use in this case.

I don't believe PommeDeTerre is saying that semantic information needs to be mixed with presentation information. He or she is asking why you have two different syntaxes when all you're doing in both cases is representing a nested tree.

For example, you could have your data structure as:

    (books (book (title "Little Schemer")))

And then define your presentation as:

    (books (book (title (:color "ff6666"))))

I'm not saying that this is necessarily the best representation, but it's more consistent and simpler for a machine to parse than XML/HSS.

Was that sarcastic? S-expressions aren't readable.

Furthermore, we used to have a mess of a language that combined style and markup. We moved onto separating the 2, so we could develop more flexible websites.

Okay, I don't get it: why does everyone assume that having the same syntax for trees automatically means you have to put everything in the same tree?

Having reread the comment, I can see what you mean. There are massive advantages to 2 different syntaxes.

What massive advantages are those? I mean there are obvious disadvantages in that you have to write two different parsers, and learn two separate syntaxes. What are the advantages?

wavejester, for the sake of not reposting:read below.

So if I understand you right, you're using XML for data because it's well known, and you've created HSS because XML isn't really suitable for styling?

Presumably that means you're prioritising developer familiarity over architectural simplicity?

No. XML is clear and precise. It supports namespace extensibility.

It also provides a very big ecosystem that AXR actually use.

I'm not certain what you disagree with. XML has a big ecosystem, but presumably you felt it wasn't suitable for use as a styling syntax, otherwise you wouldn't have used HSS?

XML is clear and precise and extensible, but it's not exactly known for having a simple syntax, right? So aren't you choosing XML primarily because of the ecosystem and tooling?

I'm not sure what your point is.

I consider XML proved itself as a reliable data layer. JSON is not considered because it does not even supports attributes and extensibility.

I would really appreciate if you would consider contributing some nice ideas. Come over to the mailing list :P

I wasn't leading to a particular point, I was just wondering what your design philosophy was.

By choosing XML you're using a syntax with a wide ecosystem and very good tooling, but you're given up some simplicity in your design. Presumably you consider this to be a good trade-off, and I don't necessarily disagree with that choice.

My concern is that in order to compete with an existing standard, a technology either needs to be a considerably better, or trivial to substitute for the old standard. XML/HSS seems to satisfy neither criteria, clinging to familiarity on the one hand, but having a high barrier to adoption on the other.

We are going to give this a second thought.

> S-expressions aren't readable.

This isn't exactly an objective truth. Personally, I find them more readable than XML or JSON.

I find them almost exactly as readable as XML, and with good highlighting, probably more so.

I'd say S-exps are definitely a lot more writeable than XML, at least.

Reply on news.ycombinator.com