What's in an API: an analogy for Verge writers
NOTE: this is not a dig at the Verge writers, although it may sound like i am poking a little fun at them. i like and respect you and your work, i am a regular reader, and i understand that you can't have specialist knowledge in every field you are covering..
it seems some Verge writers and editors don't understand all the intricacies about what's an API, how it is used, can it be copyrighted, should it be copyrightable, can you reverse-engineer it, and would you need to (you wouldn't in this case).
i didn't think that writing this article would be helpful to anyone since it seemed that covering this Goracle trial would soon come to end, but i heard Nilay say this may likely get to the supreme court, so i decided to write this down in hope it can clear some things to at least some people.
in which i compare android/dalvik to theverge.com
this occurred to me while watching the latest vergecast, and you can be offended by this analogy, or you can think about it a little. theverge.com is basically engadget.com done better. it covers basically the same topics (often the same articles), the business model is pretty similar (as far as we know), half the staff is the same (including senior editors), we have a podcast (a mobile podcast), once-a-month highly produced show, liveblogs, daily news, rumours, product reviews, and occasional op-eds and long think-peaces (i was never a regular reader of engadget, so i don't know how common they were).
also, what i can only presume is better at the verge: leaders/owners who understand the internet/blogs better than AoL.
and since i am not a great writer able to craft a great parable, i will be pretty blunt: this seems pretty similar to how google (sb nation) hired a bunch of ex-java people (Josh, Nilay, Paul...), to make a better java (better engadget), call it dalvik (the verge), without the D-list players from SUN (AoL corporate people).
was this a little shady? ex employees, en masse, going to work somewhere else to directly compete with old employer? yeah, maybe a little bit, but that happens every day in all the industries, and is mostly considered fair play (well, maybe not the en masse part, but still), and definitely not against the law.
ok, but where do the APIs come in?
this is where the analogy breaks a little, and doesn't relate to what happened at theverge anymore, but can still be useful in explaining the concept of APIs.
imagine if one of the most important features theverge.com had to have (at launch), in order to be familiar (useful) to old engadget users (java programmers) were some two dozen most important tech articles from the archives, at the same (similar) URL on the new site.
take for example, this iphone 4 review at http://www.engadget.com/2010/06/22/iphone-4-review/. you need to write another review of iphone 4, but it must satisfy these requirements:
- you must have the same URL (only on theverge.com)
- thus you need to use the same slug, and thus, the article title must be the same
- you must cover these 5 topics: hardware, software, ecosystem, design and notifications
- you must call upon the same 5 URLs as reference documentation (from the apple.com site)
- otherwise, you are free (required) to write in you own (different) words, you own opinions on the product (that probably didn't change much since 2010.. ;)
and you need to do the same for some two dozen most important articles from the last decade..
no reverse-engineering required
contrary to what some verge writers think (Nilay and Paul on the last vergecast) there was no reverse-engineering required to create a clean-room implementation of java. all the specification of how those APIs, classes and methods should behave are all documented, in very specific detail sufficient for (re)implementation in Java language and virtual machine specifications in the open, on the SUN website.
the smoking gun emails, could they afford a license, etc..
i'm not a lawyer, but neither is Andy Rubin, so his email stating that google would need a license from sun is a layman's opinion. i would bet a dollar that after that email, some google lawyer went through all the java open-source license stuff, and found a loophole that said something like "as long as we don't use the SUN's Hotspot virtual machine, and don't call what we release Java, we can work around it".
and about if google could afford the license, well of course, except for two things that SUN wanted: to be involved (bringing, in google's opinion, D-list players), and to be paid per-device license (which would prevent google from giving away android for free and disrupting the mobile industry).
IMHO, Sun/Oracle should be grateful to google for bringing Java back from the dead. i was a java programmer (on and off) for decade and a half, and java was getting boring before android. it was becoming as interesting as COBOL (programmers will understand the reference), still running half the industry, but getting deader by the minute..
so what can we conclude from this analogy? i'm a programmer and not a lawyer, so i can't claim this is following the letter of the law, but i believe it is in the spirit of it. you can't, or at least you shouldn't be able to copyright a title and structure/description of an article.
i know that writing an API requires a non-zero amount of effort (as i'm sure writing clever and punny article titles ;) but the brunt of the work lies in the actual code of the method (words of the article), and method names/params are often as trivial, utilitarian and obvious to think of as the title: "iPhone 4 review"..