Open Systems, will we ever agree on what we mean by “open”?

Just before Christmas, I stumbled on this post on Open Systems on the ESRI UK blog. My immediate response was to splutter at the irony of ESRI trying to cash in on the wave of interest in Open, who are they talk about Open? Fortunately the holiday season (and a few single malts) delayed me from an instant rant post and gave me the chance to think more about ESRI’s view on Open Systems.


Just saying something is open doesn’t mean it isn’t closed! Thx to Alan Levine

Before reading on I do suggest that you read Alan Banscomb’s post on Open Systems and maybe also look at some definition of an Open System such as this one from the Business Dictionary:

Non-proprietary system based on publicly known standard set of interfaces that allow anyone to use and communicate with any system that adheres to the same standards. Open system standards have four basis requirements(1) they must be defined fully, so that vendors can work within the same framework, (2) be stableover a reasonable length of time, so that the vendors have fixed targets to aim at, (3) they must be fully published, so that their interfaces are publicly available, and (4) they are not under the control any one firm or vendor.

And this one from PC Mag Encyclopedia

A system that allows third parties to make products that plug into or interoperate with it. For example, the PC is an open system. Although the fundamental standards are controlled by Microsoft, Intel and AMD, thousands of hardware devices and software applications are created and sold by other vendors for the PC.

For years, the term “open systems” (plural) referred to the Unix world because Unix ran in more types of computer hardware than any other operating system (combined with Linux, it still does). Contrast with closed system.

Open Systems Vs. Open Source
Open systems refers to open platforms, whereas open source refers to the software’s source code and rights regarding its redistribution. Open systems may employ open source software or proprietary software. See open source.

Open Systems Vs. Open Standards
Open systems may or may not employ open standards, the Windows PC being the prime example of an open system that is “not” an open standard (governed by a standards organization).

On the other hand, open standards do imply open systems, and the two terms are often used synonymously. However, there is absolutely no reason why an open standard could not be employed within a closed system that cannot be extended or enhanced by a third party. See open standards.

So how does the ESRI definition of an Open System stack up against these two definitions? ESRI suggest that an Open System will be characterised by “one or more of …

  • Open Standards
  • Open Source Software
  • Open Data
  • Commercial Software “…. often engineered to open standards

Not quite the same but in fairness not that far off either although they would certainly fail on item 4 of the Business Dictionary definition “not under the control any one firm or vendor“. Interesting that they chuck in Open Data as well, not sure that is really relevant but I guess getting every possible use of the word Open into the piece sort of supports the intended message, which reminds me of this post from 2012.

On rereading their post I am still chuckling at “many of our UK customers pay a modest and limited amount for the core ArcGIS components” but I suppose this is what supports “a business model that is clear and sustainable

I think they key point is that an Open System should be open to other vendors and partners to integrate with/connect to or utilise through published interfaces (whether they are formal open standards or de facto standards or just widely adopted common practice). The openness of a system is not based on whether some components are proprietary or open source, it is about design, interoperability, swappability (is that a word?) and perhaps most importantly about an attitude.  The extent to which ESRI have moved from an old model that might have been described as encouraging vendor lockin towards Open Systems is illustrated by their active campaign to be seen as Open, publishing interfaces, code packages, their endeavour to get their GeoServices REST API adopted as an OGC Standard and of course their support for several “Open” events including FOSS4G 2013.

Perhaps 2014 will be a year where procurement starts to focus more on the openness of system design and interfaces rather than the licensing of source code, that would be good for everyone working in public sector IT. Having seen a couple of tenders arrive in the last couple of weeks with specific reference to ESRI software in the requirements I think we still have a way to go.