Discussion

RELAX NG home page

jitl: the compact non-xml syntax is neat: https://relaxng.org/compact-tutorial-20030326.html#id2814005it reminds me of TypeScript.As for XML itself, it seems like it was a huge buzzword fad in the late 90s/early 2000s, but it must not have lived up to the hype or we’d actually be using it today instead of JSON and Protobuf. I got to computer programming around when the web gave up on XHTML, so i’m not really sure what to make of the XML cultural moment. The vibe i get is of focus on precise data semantics for its own sake, very Cathedral, effort that didn’t end up delivering benefit to humans. What do you think?

masklinn: I’ve become ambivalent about relaxng, I used it a bunch because I like(d) the model, and the “Compact” syntax is really quite readable, and it’s a lot simpler than XML schemas.However the error messages, at least when doing rng validation via libxml2, are absolutely useless, so when you have a schema error finding out why tends to be quite difficult. I also recall that trying to allow foreign schema content inside your document without validating it, but still validating your own schema, is a bit of a hassle.

IshKebab: I still think the reason XML failed is largely because it's a document markup language not an object serialisation language, and 99% of the time you really want the latter.You don't need attributes, you probably don't need namespaces, you probably do want at least basic types.Look at this for example: https://docs.rs/serde-xml-rs/0.8.2/serde_xml_rs/#caveatsJSON solves all of that for serialisation. The only problem with JSON is it has ended up being used for configuration, and then you really need at least comments. I wish JSON5 was as well supported as JSON is.

riffraff: XML as a document markup language was neat imvho.Like, I remember working with DocBook XML[0] and it was fine. And the idea of being able to use different namespaces in a document (think MathML and SVG in XHTML) was neat too.The problems arose from the fact that it was adopted for everything where it largely didn't make much sense. So people came to hate it because e.g. "a functional language to transform XML into other formats" is neat, but "a functional language written in XML tags" is a terrible idea"[1].Likewise, "define a configuration in XML" seems a good idea, but "a build system based on XML plus interpolation you're supposed to edit by hand" is not great[2].So people threw away all of the baby XML with the bathwater, only to keep reinventing the same things over and over, e.g. SOAP+WSDL became a hodgepodge of badly documented REST APIs, swagger yaml definitions and json schemas, plus the actual ad-hoc encoding.And I mean, it's not like SOAP+WSDL actually worked well either, it was always unreliable. And even the "mix up namespaces" idea didn't work out, cause clients never really parsed more than one thing at a time, so it was pointless (with notable small exceptions). XML-RPC[3] did work, but you still needed to have the application model somewhere else anyway.Still, JSON has seen just as much abuse as a "serialization" format which ended up abused as configuration, schema definitions, rules language... It's the circle of life.[0] https://docbook.org/[1] https://developer.mozilla.org/en-US/docs/Web/XML/XSLT[2] https://ant.apache.org/manual/using.html[3] https://en.wikipedia.org/wiki/XML-RPC

tannhaeuser: The XML spec starts like this:> The Extensible Markup Language (XML) is a subset of SGML that is completely described in this document. Its goal is to enable generic SGML to be served, received, and processed on the Web in the way that is now possible with HTML.Where "generic SGML" refers to markup beyond the basic HTML vocabulary hardcoded into browsers, such as SVG and MathML. XML was specifically designed such that mere parsing doesn't require element-specific rules such as SGML-derived HTML tag omission/inference, empty elements, and attribute shortforms, by excluding these features from the XML subset of SGML. Original SGML always required a DTD schema to inform the parser about these things that HTML has to this day, and not just for legacy reasons either ie. new elements and attributes making use of these features are introduced all the time (cf. [1]).Now XML Schema (W3C's XML schema language, and by far the most used one) isn't very beautiful, but is carefully crafted to be upwards compatible with DTDs in that it uses the same notion of automaton construction to decide admissability of content models (XSD's Unique Particle Attribution rule), rooted in SGML's zero lookahead design rationale that is also required for tag inference. Relax NG does away with this constraint, allowing a larger class of markup content models but only working with fully tagged XML markup.XML became very popular for a while and, like JSON afterwards, was misused for all kind of things: service payloads in machine-to-machine communication, configuration files, etc., but these non-use cases shouldn't be held against its design. As a markup language, while XML makes a reasonable delivery or archival language, it's a failure as an authoring language due to its rigidity/redundancy and verbosity, as is evident by the massive use of markdown and other HTML short syntaxes supported by SGML but not XML.[1]: https://sgmljs.sgml.net/docs/html5.html

imtringued: >And I mean, it's not like SOAP+WSDL actually worked well either, it was always unreliable.I don't think it ever worked. See this [0]. It's pretty crazy that people build one of the most complex and verbose data exchange formats in the world and then it turns out that duplicating the open and close tag and including the parameter name and type in the attributes bought you nothing, because implementations are treating your SOAP request as an array of strings.[0] https://snook.ca/archives/other/soap_request_pa

riffraff: [delayed]

Devasta: It is heavily used today, it's everywhere in finance and banking, it's just not on the web, so browser and web devs are forced endlessly to re implement shitty versions of things they could have had out of the box decades earlier.

vbezhenar: XML is fundamentally incompatible with commonly used programming data structures, namely lists/arrays and structs/maps. That fundamental mismatch caused a lot of friction when people use XML for data exchange between programs. JSON is clear winner here.XML is absolutely fine for data that maps naturally to it. For example text markup. While HTML technically is not XML, it's very close to it and XHTML still is a thing. Probably most people wouldn't enjoy using JSON to encode HTML pages.

tannhaeuser: > XML is fundamentally incompatible with commonly used programming data structures, namely lists/arrays and structs/maps.Another way to say this is XML is a grammar formalism that deals purely with serialisation rather than higher-level structures that might be serialised such as co-inductive data structures.> While HTML technically is not XML, it's very close to it and XHTML still is a thing.XML and HTML-as-serialisation-format are both subsets of SGML.

tclancy: Huh? I always felt some of the failure/ bad reputation of XML was how it got tortured by devs who did not understand database normalization. If you “get” 3rd form normalization, xml works fine for the relations you describe, unless I am missing something.To be clear, I am not being snide and would be interested in the cases you’re thinking of.

hnlmorg: It isn’t incompatible. It’s just a massive superset of what is needed.JSON offers simplicityYAML offers readabilityXML offers a massive feature set.For what we need 99% of the time, simplicity and/or readability is a much higher requirement.As for TOML, I honestly don’t understand why anyone likes that.

dvdkon: I don't think it's a superset. You can represent any structs-and-arrays data in XML, but you have to make non-trivial mappings to make it work.The obvious way is to use elements for everything, but then you're mapping both structs and their fields (very different concepts in e.g. C) to elements. Attributes map nicely to struct fields, but they only admit string values.

hnlmorg: That’s why it’s a superset ;)You can map anything to it but that flexibility means you then need to define schemas et al to ensure complianceThe schema thing isn’t actually unique to XML either. you can do the same thing with JSON and YAML too. But in my opinion, if you get into the realm of needing schemas in JSON then you’re at the point where you shouldn’t be using JSON any longer since you’re now violating the “simplicity” argument which is JSONs only real advantage over other formats.

lolive: Watch how dict2xml or xml2dict handle JSON to XML automatic mapping. Both format carry 99% of the same structural infos in their respective serialization.

rapnie: > it's not like SOAP+WSDL actually worked well either, it was always unreliableThis is comparable to saying that "multiplayer distributed architecture at scale" never worked well and was unreliable. All depends what your needs are and how the design and implementation satisfies them. SOAP+WSDL were part or a larger technology vision of Service Oriented Architecture (SOA) with all the complexities of distributed architecture. And the attempt was to make all of that open standards based.I worked in Print at the time at one of the largest companies (now gone bust), and can confidentally say that SOAP+WSDL worked perfectly for us, and made it way more reliable to tie all these very specialized printing equipment with archaic languages and interfaces together, increasing productivity and efficiency of the entire print process.

cess11: "XML is fundamentally incompatible with commonly used programming data structures, namely lists/arrays and structs/maps. That fundamental mismatch caused a lot of friction when people use XML for data exchange between programs. JSON is clear winner here."I'm not so sure about this. When you have a schema it becomes possible to generate your object code, and then your only immediate interface with the XML file is a rather simple instruction to unmarshal it, the rest of the time you're dabbling within your programming language of choice.

dvdkon: Eh, XML is as much of a superset of JSON as the Turing machine is a superset of context-free grammars. The former has all the _power_ of the latter and more, but the mapping between them is non-trivial, far from an embedding.I think it's fine to say C#'s data model is a superset of Java's or Rust's a superset of C's, but XML and JSON is too far for me, personally.

hnlmorg: I didn’t say XML a superset of JSON. I said it’s a superset of common programming abstractions like lists and maps.

yrashk: Relax NG Compact is a great way to communicate schemas to LLMs in a token-efficient way.