APIs that deserve more publicity

[Jplus]

Most coders hear about Qt, JQuery, MySQL or OpenGL at some point. However, there are also very good APIs out there that get much less publicity. The purpose of this thread is to inspire us by giving those APIs the credit they deserve.

I invite everyone to post suggestions. There are no restrictions except that it must be an API (in a broad meaning of the word) and that it deserves more publicity. I'll try my best to keep the opening post updated with all new additions.

Disputes are possible. If you think a certain API shouldn't be in here, let us know why and see how others respond. We'll use a temporary poll to decide on the matter when necessary.
Other suggestions for improvements of the opening post are of course also welcome.

The lists below include my own initial contributions.

By topic

Application interface

• Adobe Source Libraries (C++)

Audio, tagging

• Mutagen (Python)

Concurrency and parallelism

• Control.Parallel.Strategies (Haskell)
• FastFlow (C++)

Data interchange

• JsonT (JavaScript)
• YAML

General purpose

• Folly (C++)

Graphics, 2D vector

• Raphaël (JavaScript, DOM)

Network applications

• POCO (C++)

Regular expressions

• dk.brics.automaton (Java)

Relational database management

• PostgreSQL
• SQLite
• YQL (web services; proprietary, restrictive license)

Scripting

• LuaJIT (Lua, C, C++)

Web, browsing

• Mechanize (Python)
• WWW::Mechanize (Perl)

Web, search and data retrieval

• DuckDuckGo
• YQL (integrated, SQL-like; proprietary, restrictive license)

By language

C and C++

• Adobe Source Libraries (Application interface)
• FastFlow (Concurrency)
• Folly (General purpose)
• LuaJIT (Scripting)
• POCO (Network applications)

Haskell

• Control.Parallel.Strategies

Java

• dk.brics.automaton (Regular expressions)

JavaScript, HTML and DOM

• JsonT
• Raphaël (Vector graphics, DOM)

JSON

• DuckDuckGo (Web search)
• JsonT (JavaScript)
• YAML

Lua

• LuaJIT (C)

Perl

• WWW::Mechanize (web browsing)

Python

• Mechanize (web browsing)
• Mutagen (Audio tagging)

SQL

• PostgreSQL
• SQLite
• YQL (web data retrieval; proprietary, restrictive license)

[cemper93]

Nice thread

I don't think WebGL should be on the list. Almost every news page for programmers or non-programmers has hyped it at some point.

I would want to add the mutagen library for Python, a sophisticated audio tagging library. It's the kind of library you only need so often, but the ease of use of which is delightful when you need it.

[Jplus]

Thanks. I added Mutagen and removed WebGL.

Together we can build some great lists.

[headprogrammingczar]

The DuckDuckGo API.

Also, in the category of concurrency/parallelism: Haskell's strategies. It's a set of higher order functions for specifying how to evaluate an entire data structure in parallel.

[MHD]

The llvm is pretty great for building functions at runtime.

I wonder if there is any implementation of proper Tompson NFA Regular Expression out there? If so, put that on the list.

[EvanED]

I wonder if there is any implementation of proper Tompson NFA Regular Expression out there? If so, put that on the list.

This Java library creates DFAs from actual REs using Borozowski's derivative-based construction. (There are links to C and C# ports too.) I get the impression that Thomspon's construction is better for some things, but Borozowski's is awesome because it allows you to have things like general not and intersect operators and not have to do determinization during the RE->FA construction.

[freakish777]

Wolfram Alpha's API

[Jplus]

I added DuckDuckGo, Haskell Strategies and the Brics regex library.

@MHD: could you provide a link?

@freakish777: does Wolfram Alpha really deserve more publicity? I'm under the impression that it's rather famous.

[headprogrammingczar]

The Wolfram Alpha API is awful. It gives back way too much information with no obvious structure. The only useful thing you can do with any kind of ease is just present the data the same way WA does.

[freakish777]

@freakish777: does Wolfram Alpha really deserve more publicity? I'm under the impression that it's rather famous.

You said APIs, not the companies that made them, which leads me to:

The Wolfram Alpha API is awful. It gives back way too much information with no obvious structure. The only useful thing you can do with any kind of ease is just present the data the same way WA does.

If it had more publicity and got more people trying to use it (who also complained), hopefully they'd make it more useful (like giving formatted results).

I mean, everyone knows Google has APIs, but does everyone know Wolfram Alpha does as well? If so, then yeah, it doesn't need more publicity. If not, then I don't see how famous the company is (as compared to the API) should factor in.

Basically, I think this list shouldn't just be "Which relatively unknown APIs do useful things" but also "or could do useful things in the future if they got more exposure, based on what we know about the company that supports it." I mean, I'd love to see Google's APIs be more useful, but they're fairly well know, so adding to their publicity probably won't solve that problem.

Does that make sense?

[Jplus]

I kind of see your point, but for as far as I'm concerned this list is about displaying APIs that are very useful (preferably good) already, not about getting APIs a larger user base so that they might perhaps become more useful in the future. headprogrammingczar made quite a strong point that the Wolfram Alpha is probably actually not so useful.

Also, I actually meant the API and I'm still sceptical about the notion that they might be receiving too little publicity. For starters, the only people who responded to your suggestion so far turned out to know about it, and I heard about the API several times before.

Secondly, the very fact that the API is provided by a well-known company that makes a service out of providing information on the internet should ring a bell (cf Google, Twitter, Yahoo). Compare this with Adobe and DuckDuckGo (the only other commercial companies currently in the list). Adobe isn't providing any information services, and since their applications are proprietary it isn't so obvious to expect that they provide an open source API. DuckDuckGo does provide an information service, but currently they're not widely known yet; when that changes, the API should probably be removed from our list.

Finally, if an API is less useful then that automatically means it deserves less publicity.

However, if you insist, or if somebody else kicks in to support the inclusion of Wolfram Alpha, we'll set up a poll and let the majority decide.

[freakish777]

However, if you insist, or if somebody else kicks in to support the inclusion of Wolfram Alpha, we'll set up a poll and let the majority decide.

I do not insist, just personally would like to see more people leverage APIs to force companies to better support them (especially when that company could be making their API incredibly useful, such as in this case).

[Jplus]

*bump*

Mechanize has been mentioned a while ago elsewhere on these forums. It's a nice Python library for browsing the web, derived from the Perl module WWW::Mechanize. I guess I'll put both on the lists?

Did anyone else come across interesting stuff in the meanwhile? I bet there's so much more to discover.

[D-503]

YQL can do some really cool stuff.

[Jplus]

YQL looks really interesting indeed. Contrary to the other APIs in the list it's proprietary with a restrictive license, do we consider that a problem?

On an unrelated note, I realized that if PostgreSQL is on the list, so should SQLite.

(Now going to add Mechanize, WWW::Mechanize, YQL and SQLite.)

[Jplus]

POCO is a nice C++ library for network applications. Folly is Facebook's general purpose collection of tidbits, also for C++.