Short cuts 56

by Paul Stephens

Paul Stephens takes a sideways look at the world of IT.

HardCopy Issue: 56 | Published: May 1, 2012

The Trial of the Century (part 1)

Short Cuts was thrilled to see our former favourite CEO, Jonathan ‘Dolphins’ Schwartz, back in the thick of it in April as a witness in the Trial of the Century between Oracle and Google. The database giant is suing the search colossus over the use of a programming language in a mobile phone OS (which says something about the diversification of both companies) and Schwartzy, the Sun Microsystems boss who was left ‘without a role’ when Oracle bought the company (and with it the Java language which all the fuss is about), appeared, to no-one’s surprise, as a witness for the defence.

Jonathan Schwartz

Former Sun CEO Jonathan Schwartz – back in the thick of it as a witness in the Oracle-Google Trial of the Century.

Oracle was claiming that although the Java language is free under a GPL, its API isn’t, and that by giving its Android Java implementation API elements with the same ‘structure, sequence and organisation’ as Oracle’s (even if the actual code is different), Google violated copyright big time ($6 billion big, in fact, when combined with a separate patent issue, although this was revised to a mere $1 billion when the judge advised them to tone it down).

This could have big implications for the software industry, since in languages like Java the bulk of functionality is in the API, leaving an open source language without one a bit like a one-legged racehorse. Oracle’s definition of the Java API includes the java.lang package which “provides classes that are fundamental to the design of the Java programming language”, so it seems its idea of copyrighted functionality starts pretty much straight after curly braces and the ++ operator.

Open Source Superhero Schwartzy’s performance in the witness box may not have been exactly what Google wanted, although it did help a bit. “We didn’t like it, but we weren’t going to stop it by complaining about it,” he said, referring to Google’s go-it-alone Java implementation, although as Google’s lawyers pointed out he’d perked up enough later to offer his “heartfelt congratulations” to Android’s authors in a posting on his (in)famous blog.

Oracle countered by wheeling out former Sun Chairman Scott McNealy to say that Schwartzy had got it all wrong (plus, rather cruelly, that he never read his blog), and asking the Judge to ban Schwartz from saying any more unhelpful stuff. Google retaliated by asking Oracle witness Tim Lindholm, a Sun-then-Google Java engineer who allegedly Knew Too Much, to be banned as well, and from there it all got a bit chaotic.

The judge eventually sent the jury out to decide on the copyright issue (despite not having ruled yet on whether APIs could actually be copyrighted), and they duly returned with a split verdict – yes, Google had violated Oracle’s copyright on the Java API (assuming such copyright exists), but no, they couldn’t decide whether Google’s actions amounted to ‘fair use’, a legal term which apparently means “they nicked it, but then you do, don’t you?”

This left both sides claiming victory, and a nervous software industry still not sure whether the class name ‘Byte’ has the same copyright protection as, say, ‘Harry Potter’. The good news (if you’re a San Franciso-based lawyer) was that this was just Round One: as HardCopy went to press, they were starting Round Two, to decide whether Google violated Oracles’s patents on the Java VM. Short Cuts naturally hopes that Justice and Fair Play will be the winners, but we’re not sure how we’ll know.


Lotsa Bytes

Wikipedia's Multiples of bytes table

Wikipedia’s table of byte multiple names only went as far as the Yottabyte, but those byte multiple naming guys were already on to the Brontobyte and Short Cuts can reveal that there’s more to come…

Reading round the Web about Red Hat Storage 2.0 (covered in the News on page 4), we were impressed to find that the mega filing system has scalability up to a whopping 72 Brontobytes of data, each one equivalent to 1,024 Yottabytes, and each one of those a lot of bytes in its own right (280 to be exact).

We’d never heard of either number though, and when we found that our industry-standard reference source on byte multiple naming (Wikipedia) hadn’t heard of the Brontobyte either, we began to suspect that those multiple-naming guys were making it up as they go. In fact they’ve only just got started, as our roundup of forthcoming multiple names (currently in beta but due to ship in Q3) shows:

Larribyte: Named after Oracle chief Larry Ellison’s annual income, including share options. Described by a leading mathematician as “an unimaginably large number.”

Agilobyte: Named after the average number of updates issued per quarter for software developed using Agile Development Methodology. Can vary depending on how well thought-out the previous quarter’s updates were.

AppleBite: Named after the profit margin achieved by Apple on its iPhone and iPad devices. Described by a Nokia executive as “an unimaginably large percentage.”

Adobibyte: Named after the number of applications in the latest Creative Suite release, including the ones whose purpose isn’t entirely clear but are something to do with workflow. Described by a leading mathematician as “not quite as big as a Larribyte, but getting there”. Not to be confused with the even bigger…

Addonobyte: named after the number of ‘free’ apps in the Google Play store which are actually just there to persuade your kids to buy expensive add-ons without telling you.

Patentobyte: Named after the number of patent infringement lawsuits Microsoft, Oracle, Google, Apple and Samsung have launched against each other in the past four quarters. Not widely available, as its use to describe this number has been patented by Adobe.

Ballmerobyte: (that’s enough bytes – ed)