Hacker News Reader

Comment by roenxi | Hacker News Reader

The SQL Standard appears to me to be some sort of political tool, likely developed and used by Oracle. It is the most significant standard I know of that isn't publicly available.

Seriously, if you want a legal copy that costs something like 200 Swiss francs. This isn't how serious comp-sci standards work (the TCP people just throw out RFCs). Especially since the SQL standards committee probably don't know how to design a language given what most SQL implementations look like. I can't be sure though, I'll probably never be in a position to read the official standard.

Reflect on the madness. Postgres? Free. SQL Standard? 200 francs. Postgres makes a better standard than the SQL standard. The ISO standard thing isn't here to support people using or building SQL databases.

View on HN

I sometimes fail to understand the disconnect in some of us here: you want six figure salaries, AND have standardization committees work for free (and all software and AI models should be provided for free as well).

This would all make sense if you were in favor of fat governments, but that does not seem to be the case either.

The entitlement baffles me. Or am I accidentally combining opinions of different people and making a misjudgment?

I'm not asking for anything; the postgres documentation is more important to me than the standard and already free. I'm pointing out that the standard isn't intended for implementers or users. By ruling out options, my guess is it is intended for governments to select Oracle DBs rather than alternate (OSS?) DBs - or just to put petty impediments in the way of new DBs trying to get started. Otherwise why is the standard the only component of the entire chain that is not free on the internet? It would be bizarre if ONLY the standards committee needs $$$ to operate. They have the easiest job and there are natural incentives for the different SQL implemeters to cooperate. Why else would the core postgres maintainers be expected to pay 200 francs for the privilege of getting a copy of a standard that they are implementing for free? That way lies madness.

And we know the standards committee would be willing to work for free or people are available who would do the work (indeed, we can probably recover the standard from the postgres docs since they mention the things that aren't standards compliant - https://www.postgresql.org/docs/current/features.html). The postgres maintainers do notably more work than the standards committee for free. How hard is it for them and Dr Hipp to throw out a document saying where the common ground is? It is far less work than they've already put in to the free docs. The useful parts SQL standard is an appendix compared to https://www.sqlite.org/docs.html

At least, I assume. As previously mentioned, I've never read the standard and as far as I can tell nobody is allowed to link it for me.

EDIT I feel like I should link some standard here, so in lieu of the SQL standard which I can't link, here is the TCP standard: https://www.ietf.org/rfc/rfc9293.html

It is entirely possible for people to be paid to produce something that is made available at no cost to certain other people. For example, that hand sanitizer dispensor at the pet store. Yes, somebody is paying for it.

In the case of a standards body, the people that do the work can get paid in various ways. The finished product can be published "for free".

In some research contexts this could very much be expected. If researchers were paid through tax-funded grants, for instance, the results of the research should be available to those tax payers. If they are also available to non-taxpayers, perhaps there is some altruism at play.

It's not an absolute given. But it's not an unreasonable expectation.

Of the top of my head, there are also:

- ISO 8601: dates and times

- ISO/IEC 9899: the C standard

- ISO/IEC 14882: the C++ standard

> This isn't how serious comp-sci standards work (the TCP people just throw out RFCs).

This isn't how most _web_ standards work.

True, although for both C and C++, standards drafts are prepared on GitHub. At least for C (but I presume also for C++) you can check out a revision which is functionally equivalent to the ISO standard (but minus the branding).

FWIW, the PDF spec also costs some crazy amount like that for something so ubiquitous.

As of 9 days ago, the PDF 2.0 spec has become freely available: https://www.pdfa.org/announcing-no-cost-access-to-iso-32000-...

It's how most standards work outside of comp-sci though.

Reply on news.ycombinator.com