Re: [xml-dev] It's too late to improve XML ... lessons learned?

(The excellent Stephen and I are not agreeing much so far this thread! It cannot last.) I don't see most of his last points:

Apart from _javascript_ most every different host language has had its own JSON parsers. There isn't just a single pre-existing library every language and platform used. So it did need a development effort.

And JSON was notorious for parser incompatibilities on the edges, for its first decade at least. It needed bug fixes, which means agreement, which means a community caring about agreeing (rather than winning), which sometimes is helped by a formal process and document (standards), to some extent. So it did need agitators for interoperability = conformance.

And there was a standard for _javascript_: the first edition of EcmaScript came out in 1997. So to the extent that JSON was faithful to EcmaScript (3rd ed) it is a profile of an existing standard. (And JSON made its way up the formal standards ladder steadily: by RFC, by ECMA, by ISO. So it did have standards, even if as a confidence booster that JSON is a thing not a fad.

Rick

On Wed, 5 Jan 2022, 11:35 pm Stephen D Green, <stephengreenubl@gmail.com> wrote:

> And how can JSON not count? If something does not have drivers for its success, surely it would not have succeeded?

I think JSON succeeded because it was already existing and got discovered so it was free and already had traction and some sanction as part of _javascript_. So it did not need the ingredients to success that a new technology would need. It did not need a vacuum to fill, developer adoption, new parsers, bug fixing, security patching and rethinks, lengthy specs, standardisation, expensive consultants, evangelists. These were hurdles XML had to overcome and a new technology would have to overcome.

On Wed, 5 Jan 2022 at 10:53, Rick Jelliffe <rjelliffe@allette.com.au> wrote:

On Wed, Jan 5, 2022 at 10:55 AM Stephen D Green <stephengreenubl@gmail.com> wrote:
"Now that XML exists, there is no driver for anything better. JSON was a happy accident so that does not count."

Your kidding, right?

I don't think we reached a steady state of technology in 1997, and nothing has changed in the subsequent 25 years that could possibly make it worthwhile to revisit the goals and engineering decisions of XML? No Big Data? No 3G/4G/5G? No GPUs, cheap SSDs, SSE/AVX, etc? No ubiquitous Unicode? No HTML 5? No cloud computing? No ubiquity of XML-in-ZIP formats? No Git? No failure of Web Services? No ubiquitous availability of interpreted languages with eval() functions? No REDIS? No XSD and its disappointments? Etc. etc.

And how can JSON not count? If something does not have drivers for its success, surely it would not have succeeded?

It seems that many Big Data systems read in a dialect of JSON, with one JSON "file" per line. They scan for the appropriate line then read in the record. How is that not an example of the kind of non-ephemeral data that we thought XML should be good at? XML at W3C stopped responding to drivers; diverted first by giantism of the data-binding crowd then by the dwarfism of the sugar-free crowd.

(What JSON showed was that the driver was for a header-free markup language with richer delimiters that allowed direct specification of datatypes through syntax, trivially converted to conventional datatypes and structures: not richer but more complex, nor simpler but less rich, but rich and simple: more expressive syntax.)

A standard can breed or it can go into palliative care.

Rick

--
----
Stephen D Green