Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

There are over one trillion SQLite databases in active use

sqlite.org

Post reply on HN
Print view View original

81–90 of 92 posts

Re: There are over one trillion SQLite databases in active use

#81

Post by detaro »

Taywee wrote: post #80
detaro wrote: post #79

Earlier quoted context omitted.

It's sort-of misleading IMHO. A shared backend would be fine (i.e. I believe the independent IndexedDB implementations all end up putting stuff into sqlite, but nevertheless count as independent since they demonstrate that the API isn't tied to it), but the spec draft said > User agents must implement the SQL dialect supported by Sqlite 3.6.19. as the only definition on what's allowed SQL, which is a no-go for a spec…

Ah, I see. The wording on the Wikipedia page isn't as good as the wording on the specification draft itself: > This document was on the W3C Recommendation track but specification work has stopped. The specification reached an impasse: all interested implementors have used the same SQL backend (Sqlite), but we need multiple independent implementations to proceed along a standardisation path. It seems like they didn't…

It would likely be necessary to have an independent interpreter for the SQL dialect at least, yes. I.e. I assume that something that parses the SQL from scratch and then down the line hands stuff to sqlite as the storage engine would qualify, since the SQL dialect is the point of contention. (The "multiple independent implementations" requirement is afaik also not 100% hard - if everyone agrees to ignore it, it can be ignored, but it's a strong objection if someone disagrees)

Re: There are over one trillion SQLite databases in active use

#82

Post by wodenokoto »

trasz wrote: post #45
thepasswordis wrote: post #15

SQLlite has also got probably the most well-tested code of conduct of any open source project in the world (aside from the other open source projects which use this CoC) https://sqlite.org/codeofethics.html

Isn’t SQLite essentially not accepting outside contributions at all? In such a case this CoC really might work just fine. As would a red “GO AWAY” banner.

> This document continues to be used for its original purpose - providing a reference to fill in the "code of conduct" box on supplier registration forms.

Re: There are over one trillion SQLite databases in active use

#83

Post by dang »

sdfghderwg wrote: post #73
dang wrote: post #63

Earlier quoted context omitted.

Let's not get distracted by an extraneous provocation. As you can imagine, this has come up frequently when people run into that bit, and tends to lead to the same discussion over and over (and ultimately, religious flamewar). There's a new site guideline about this actually: " Please don't pick the most provocative thing in an article and rush to the thread to complain about it. Find something interesting to comment…

> "Please don't pick the most provocative thing in an article and rush to the thread to complain about it. Find something interesting to comment about instead." So you agree it's a provocative thing. Good, that's all I meant to convey.

Yes, but we're specifically asking you not to convey it, for the same reason that one doesn't convey lit matches to dry forests.

Re: There are over one trillion SQLite databases in active use

#84

Post by jrimbault »

jrimbault wrote: post #52
rpmisms wrote: post #48

Earlier quoted context omitted.

I don't think anyone can reasonably get angry over someone else following these. Good rules.

> Speak no useless words or words that move to laughter. I'm extremely biased coming from a family with a long tradition of atheism and even anticlerical "militantism".

PS: the second line here, though true, is meant as a meta-joke contradicting the quoted commandment

Re: There are over one trillion SQLite databases in active use

#85

Post by literallyaduck »

claytongulick wrote: post #2

They'd have one trillion + 1 if they had better JSON support. I'd be using it in a large healthcare project I'm working on now. Unfortunately, the JSON support is basically "serialize to a string". Postgres is miles ahead with jsonb. Shame, I <3 sqlite.

Sqlite is just that, lightweight. If you want better JSON support specialized for what you want, extend it. The product allows extensions easily.

Re: There are over one trillion SQLite databases in active use

#86

Post by ec109685 »

Taywee wrote: post #78
Someone wrote: post #72

Earlier quoted context omitted.

I’m fairly sure that refers to https://en.wikipedia.org/wiki/Web_SQL_Database : “Web SQL Database is a deprecated web browser API specification for storing data in databases that can be queried using SQL variant. The API is supported by Google Chrome, Opera, and the Android Browser. The W3C Web Applications Working Group ceased working on the specification in November 2010, citing a lack of independent implementation…

I see, thanks. I don't totally agree, though. The comment implied that SQLite's test suite being proprietary is what "kept SQLite out of the browser" by preventing forks, when what kept SQLite out of the browser was just that there were no other high-quality embeddable SQL libraries. Perhaps the test suite being non-proprietary would have allowed for easier forking, but the lack of an available open test suite certai…

They don't allow pull requests could be one reason. E.g. if you wanted to take SQLite in a different direction.

https://www.sqlite.org/copyright.html

Re: There are over one trillion SQLite databases in active use

#87

Post by lifeisstillgood »

thepasswordis wrote: post #15

SQLlite has also got probably the most well-tested code of conduct of any open source project in the world (aside from the other open source projects which use this CoC) https://sqlite.org/codeofethics.html

> Be in dread of hell.

I have thought about this one a bit, as an Atheist, and frankly, I have a problem with it. Ok, heaven, yes. Eternal peace and love, sounds great. Hell, obviously bad, the torture, the pain. But, to suffer pain I need to be conscious, awake, aware.

For all Eternity.

And surely the Demons must take breaks. nip out for a quick fag. Maybe a long lunch at the pub. I mean they might not tell Satan they were doing it - these are evil beings after all, not known for following the rules. So 15 minutes here, an hour there. Over a week it will add up.

And weeks add up over all of eternity for infinite time.

I mean. You could manage. If you knew you had infinite time, to think, to be. Compared to my belief, extinction, Hell is probably good runners up medal.

Re: There are over one trillion SQLite databases in active use

#88

Post by tonyedgecombe »

Slartie wrote: post #67
_xnmw wrote: post #44

One of the examples of wildly successful "one-man-army" software (conceived of and built by a single person). See also: Redis, Linux, Doom/Quake, Steve Wozniak, jQuery, Python and countless obscure modern SaaS products.

Steve Wozniak was a software project?

>conceived of and built by a single person

Clearly Wozniak is Jesus.

Re: There are over one trillion SQLite databases in active use

#90

Post by claytongulick »

acatton wrote: post #12
claytongulick wrote: post #2

They'd have one trillion + 1 if they had better JSON support. I'd be using it in a large healthcare project I'm working on now. Unfortunately, the JSON support is basically "serialize to a string". Postgres is miles ahead with jsonb. Shame, I <3 sqlite.

I love PostgreSQL, and would recommend it anytime. Even though jsonb is very powerful, in general if you store JSON in a SQL database you're very likely doing something wrong. (You're breaking first normal form) But if you really wanted to use SQLite to store JSON à la jsonb in PostgreSQL you can use generated fields[1] sqlite> create table t(id integer primary key autoincrement, data text); sqlite> insert into t(dat…

Database optimization and the level of normalization is a complex topic that varies widely based on application and workload.

JSON is absolutely a valid approach to data storage when dealing with certain data structures. As is using a relational database and jsonb data types for it. Cherry picking columns gives you the advantages for both a NoSQL document storage database as well as a relational analytics database.

This is a pretty common modern technique that Postgres, for example, has made easy.

Go try modeling a FHIR database structure using standard normalization rules.

You'll quickly discover why no one does it. Even HAPI FHIR server (the most full featured and popular FHIR server) written in Java doesn't attempt it.

Post reply on HN