Tag Archives: Oracle

Copyright on APIs and header files?

[US] Supreme Court won’t weigh in on Oracle-Google API copyright battle (Ars Technica) – Lower court to decide if API utilization is copyright infringement or fair use.

The dispute centers on Google copying names, declarations, and header lines of the Java APIs in Android.

This is such a bad idea.  APIs (Application Programming Interfaces) are fundamental to interoperability – they describe how two programs, or a program and a function library, or a program/library and an operating system kernel, or two distant computers, can interact.

Years ago some people were trying to make an argument that communication through a published protocol (remote API) constituted “linking” – that is, an application would be regarded as being linked with (for instance) a database server through the mere act of talking to it. Linking in programming is generally combining an application with libraries in a single package, and in licensing it has a special meaning. The GPL (GNU General Public License) and others “kick in” when a program is linked – that is, any linked components must be compatible with the license.

To give you an idea of how stuffed the idea of regarding protocol/API communications as linking is, it would mean that a web browser (say Internet Explorer) is linked to Apache web server through the HTTP protocol. Aside from becoming a licensing quagmire, it was just beyond daft.

The reason I bring this up is that essentially, the Oracle lawsuit tries the same again, via a slightly different route. The results are very similar.

Note that of course this is just a legal case in USA, but I do hope that the lower court finds that it’s at least within “fair use”. It could still be problematic because other countries don’t have equivalent fair-use built in to their copyright legislation. Thus, if Oracle (or others) were to pursue this same topic in other jurisdictions, it’d be very messy.

The concept of an API is not specific to computer programs. Companies have business processes (for interacting with other companies and individuals), which perform exactly the same tasks. Think of the flow of a quote, purchase order, delivery, invoice, etc. Would it be sensible to copyright an “implementation” of that, and thus potentially see someone decide whether you can use it, and if so, for what fee?

There are standards and arrangements for almost everything around us – letters fit in mailboxes, cars fit on our roads and parking lots. Those are both interesting examples because we know that some cases don’t adhere to these boundaries and it causes trouble. A car can be way too big for common parking spots.

In a nutshell, I reckon that if this were to be upheld, the law needs to be adjusted. It’s neither practical, nor beneficial to our society or economy.

Outsourcing your Public Presence

Many companies and organisations, large and small, now only interact with people through sites like Facebook and Google+. That is, no longer do they have blogs and other items on their own site that allow comments and other forms of interaction. For example, one of my local radio stations (ABC 612 Brisbane) posts links and items on their Facebook page, and while I can still send them an email, essentially all the interaction with the program makers and other listeners happens in the Facebook threads. I find this a very troubling development.

Let’s look at what the drivers are. Very simply, an external service provides convenience with less internal effort, so it saves money (many companies don’t have any web skills internally at all). Users can share posts easily with their connections, and thus an organisation possibly gets wider exposure. I understand all that, but what are the consequences of this development?

As a sideline – there’s also a perceived advantage for the user, and that is that they’re not handing over their name/email and possibly other information to so many companies – instead it’s concentrated with a few, and that makes users feel more comfortable and in control. Whether they are actually (in control) remains to be seen, as it provides information behemoths like Google and Facebook with even more ways to profile people and tie things together. What are the long-term consequences of that?

While in the case of Google+ you can get your data out, that only applies to a fairly limited recent range of time. Last time I tried, I could only export my items from the last 60 days or so. Posts prior to that were actually completely gone. Not the kind of public record a company looks for, is it? Historical threads can be quite important, particularly when people search for topics that you’ve been involved with. Even a long time later, such links provide an important source of traffic and potential business, and also support the ongoing public image. Facebook basically just owns your data, but at least it doesn’t have a limited lifetime. That is, as long as that’s what Facebook decides to do with it. Either way, you’ve handed over a lot of control, in return for some convenience, but definitely with negative consequences for you and those add up over time, even if the provider does not disappear or decides to do things with your data that are not to your benefit.

Are there alternatives? Disqus appears to be doing well. It takes care of the registration/login tasks and handles the comment process for posts, but it integrates with an existing website or blog system. I do like this approach better as it doesn’t shift your existence (in terms of the URL where users go) to an external site, but nevertheless there’s still a risk, the provider can go away at some point. Where is your data (user contacts, comment threads), can you still access it, can you get it out, and then can you do something sensible with it to not loose this rather valuable historical asset?

While pondering this article I thought of Greg Gianforte, founder of RightNow technologies and author of the book “Bootstrapping Your Business” which Upstarta gladly recommends as many of the concepts are closely aligned. I’ve met Greg when I organised a conference in the US, his story is pretty interesting. He bootstrapped RightNow in 1997 and the company went public in 2004. What RightNow has been doing was quite pioneering, they allowed companies to outsource much of their customer interaction (support tracking system, online helpdesk, feedback forms, etc) while keeping the original branding. So it was a hosted, branded, service. As you can see, Disqus is doing something similar, but just for comments. While not many of the public would have heard about RightNow, they did really well and helped many companies deliver customer service – a typical business-to-business service. We’re not talking small fry, by the way: huge multinationals such as Proctor&Gamble use them. Their current range of services is extensive, so you can just imagine how they can really enable companies, but… at the same time really tie them to this service.

Late 2011 the company was acquired by Oracle Corp. What will happen now? Oracle has bought out many other players in the Enterprise Resource Planning (ERP) and Customer Relations Management (CRM) sphere. Typically the clients get absorbed, and the original products disappear. Naturally the exact flow differs per company and product/service. However this ends up though, will it be good for the clients? I doubt it. They had a service they liked, and now they’re going to have to adapt to a new environment, or leave.  Do they even have that freedom? Imagine the huge cost of migrating to another provider (with similar risks) or taking the services in-house. But who pays for the changes if they stay? Business processes would need to change either way. I realise it is a very rough picture that I’m drawing, but the crux is this: these businesses are now highly dependent on the external service, so the extent that they get high costs forced upon them in order to survive – and if they can’t afford it, bye-bye company.

This flow of events is entirely typical, companies make business decisions that are entirely valid at the time. They outsource tasks that are not part of their “core competencies”. Even looking towards their future, with some attention to the possible risks, the decisions appear to make sense at the time. Yet the companies get bitten later, so something  has happened and that is the kind of thing that Upstarta loves to investigate. Were the choices correct, were the risks fully understood, and the consequences considered? It keeps happening to big seemingly smart companies, in different contexts (Dell computers getting slowly eaten by AsusTek: first they produced some components for Dell, then mainboards, then complete systems. And then it went directly to market with its wares, bypassing Dell).

In case of Dell/AsusTek it’s a typical example of disruptive business, but it’s interesting as well as disconcerting to observe this in the context of these critical services such as a company’s interaction with its own clients. Is that, perhaps, too much risk to take?