2. Risk: Our landscape is changing so fast there is an extreme amount of exposure to dumping resources into a solution that isn’t supported. If any one component of the enterprise stack changes you’re vulnerable and I trust those I’m paying to cover my maintenance than I do an ethereal community. When it comes to supporting my clientele I need tangible support resources. How good are the support resources for open source solutions? Is there comprehensive up toI think this article at InformationWeek (number three in the Google search for “open source support”) sums it up well. There is not one answer, there are a number of answers, and you need to choose the one(s) that make sense for your needs.date documentation? Can I call someone in an a oh $*!&% moment?
- Product support, from specialist companies with expertise in particular components (Refractions for PostGIS, DM Solutions for Mapserver, TOPP for Geoserver)
- Stack support, from generalist companies putting together mixes of components (Wheregroup)
- Community support, much maligned, but better than the technical support provided with most proprietary packages
- Training, from folks like the companies above, or specialist trainers like Open Technology Group
- Hiring project developers, an often unappreciated source of top notch trouble shooting and knowledge,
- Consultants, who have to know the tools they use
CIOs would be well-advised not to buck the open-source trend. On the contrary, they should assume responsibility for open-source initiatives and ensure that their companies have the right support structures in place... They'll find the more mature an open-source software project, the more mature support options its users enjoy.Open source requires you to assume responsibility, which is hard for an organization man, with years of CYA behind him, to do. In exchange for taking responsibility for your own infrastructure, you are rewarded with a software ecosystem where there is more than one source of support.
What do open source organizations do when their support provider isn't up to snuff? They get a new support provider.
What do ESRI customers do when ESRI support isn't up to snuff? They bitch about it on James Fee's blog.
2 comments:
Paul,
Some very good points.
Support is not some linus blanket that absolves one of thinking through a problem oneself. Your point about taking responsibility is critical to successful resolution of any issue. Also, experience with any product brings with it a serendipitious insight that is part of problem resolution.
Could one say that it is a universal truism that support for most software products - whether open source or proprietary - occurs in forums in cyberspace? Within those forums user-user/user-vendor transactions occur without money changing hands. Sure, some vendors may restrict access to registered users of the software, but these users may not even have a current support/maintenance contract. The democracy of the internet is such that where restrictions are placed (not just the right to criticise the product) after-market forums get created: they are everywhere in the IT world (eg lots of unofficial Oracle DBA and developer forums) so why should GIS be any different or special? What is more critical to the success of these forums is their flexiblity of access modes (forum vs email list etc) and their ability to attract self-motivated people who will help in problem resolution. Proprietary software has no specific advantage here.
To be honest having been at the blunt (customer) end of vendor support/maintenance fees (where they could not be split) over the years I would assert that the return on investment for the customer forced to pay the fee is very, very low. In the internet age maintenance/support fees appear more cash flow for funding development that providing real support.
Indeed, Simon, support cash is the life-blood of many companies. I would feel better about the money flowing into proprietary software companies if more of it went to, say, software.
But check out Oracle's Annual Report, page 68.
R&D expenses are actually less than sales and marketing (S&M, for short). Oracle spends 15% of it's software licensing and support revenue on actually improving the software. It spends 23% on selling that software. Proprietary software companies need to be understood for what they are optimized to be: software sales organizations. The production of actual software is merely an incidental byproduct of their core mission.
With respect to you comment on the quality of support, page 41 is instructive. They took in $8.3B in support revenue. It cost them just $1.3B to provide the support. For a profit margin of, wait for it, 84%. Larry needs a new yacht.
Post a Comment