Closed source vendors hijack the term "open"

Latest software sales pitch is 'open source by association' (not by license).

News flash to old-school software vendors: an "API" doesn't make your product "open" and it certainly doesn't make it "open source." For the second time in two days, I've seen product marketing claiming a product was open because the vendor supplies an application program interface. Phooey.

we're open
An API lets third-parties write add-ons and if you are going to release closed-source apps, an API is better than nothing. But this isn't the early 1990's. You can't legitimately say an API makes your software "open" by today's standards.

Products are open source only if users of that product can see the source of the product itself. And, to my way of thinking, a product is REALLY only open source if it is also covered by a certified open source license. And vendors that REALLY want to sell their product on the merits of it being "open" and "open source" should offer a license that is certified by the Open Source Initiative. Better still, the OSI has certified several dozen licenses. Want to be open? Choose one that is widely accepted (i.e. BSD, GPL, Eclipse).

The press release that got my ire up this morning was issued by General Telecom, which released some additional features to its RouteNGN software, a software router for carriers that helps them send TDM and VoIP traffic over the most cost-efficient routes (so General Telecom says). What got my attention was General Telecom's claims that the tool combined "a modular approach with open source development."

I thought, "Hm, maybe this is an open source router-like thing-y for carrier class networks something like what Vyatta is for SMBs and branch offices." (Thing-y being the technical term.)

But, upon closer inspection, turns out that RouteNGN is not, itself, open source. It has an API which General Telecom reportedly posted on GitHub. Apparently, it wants to make its RouteNGN tool considered open source by association, as the press release proudly proclaims that the API "is published on GitHub, a prominent hosted repository service for software projects such as Ruby on Rails and the Linux kernel."

The ultimate irony? What is really posted on GitHub, at least as far as 15 minutes of searching would reveal, is not even the API itself, but merely a wrapper written Ruby for the API.

Open source? NOT.

The second instance of attempted-open-source-by-association was for a new software/hardware/storage bundle I can't tell you about until Monday. On Monday, a vendor will be introducing what it says is the first "open" product for <ok, I can't tell you that yet>. The vendor has determined that its new product is "open" because it will be publishing an API.

Open source? NOT.

Open? Well, let me just note that Windows has had an API for decades (it is an operating system after all). After so many rounds with antitrust litigation in both in the U.S. and in Europe, I doubt anyone would call it open.

Can a software product that publishes its API legitimately claim that it is open?

Please visit the Open Source Subnet home page for more news, blogs and podcasts.

Subscribe to all Open Source Subnet bloggers

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.
Now read: Getting grounded in IoT