As Microsoft’s Tech-Ed, the biggest developer event on the calendar, takes place in Auckland this week, open source software is unlikely to be much talked about. That’s because developers divide pretty much into two camps: those who see benefits in sharing code among a community of programmers and those, exemplified by Microsoft, who believe in holding firmly to code ownership. Computerworld invited Peter Harrison (see Open up to opportunity), an open source advocate, and Tony Stewart (see below), a proprietary software believer, to argue their corners.There’s been plenty of debate over open source versus proprietary, predominantly in the shape of Linux versus Microsoft, but the real question is: what makes software “open”?
Everyone is blinded by the fact that the Linux operating system is “free”. Microsoft is no longer the low-cost option. When Microsoft came on to the scene, the other vendors said service was their competitive edge. Now it is Microsoft’s turn.
As a Microsoft developer, the open source world looks to me like a whole lot of disjointed products where you have to pick which ones you can be an expert in. If you choose right, you get support from the community that evolves; if you choose wrong you are on your own.
An open source environment can more rapidly become a unique conglomeration of bits of customised code and utilities. Organisations may not realise that they can be locked in to a company or individuals with the only true understanding of their unique implementation.
As a Microsoft developer, we have a more prescriptive platform to build upon. Our lack of access to the source code, and use of best practice, provides our clients with more certainty that what we have built is supportable by another Microsoft certified partner.
So how open is open? Is it open because you can get open source code, customise it, then wind up with your own unique version of software, or is it more open standardising on a widely used platform were everyone runs the same code?
The bottom line is that it is much easier to implement, manage and upgrade applications developed on a unified platform.
Ironically, the more people who adopt Linux and open source, the more demanding they will become in terms of R&D and support. The model pushes the support out to the consultants but clients will eventually find themselves turning to value-added services from larger vendors who were traditionally considered proprietary.
With the advent of Microsoft’s .Net initiatives, we get a first-rate environment for developing enterprise-scale applications. We get a set of servers and development languages that are not only all designed to work with each other but have a standards-based interface that allows them to work with anyone else.
By developing on a platform from one vendor, everything has been tested together; you get documentation and standardisation on how to deploy in that environment. For a development house like ours, working with one vendor reduces the cost of doing business.
If something goes really wrong, we have the supplier of all the software working 24 hours a day with us until the problem is solved.
With open source, we see some organisations playing with it where they have strong in-house development teams. These teams are able to build a degree of self-sufficiency; they are able to replace their investment in licence fees with an investment in an internal support structure.
We worked for 170 organisations in the past year and not one of them has switched to open source.
We don’t see Microsoft shops go to tender for a solution, then have someone from left-field convert them to open source. Conversely, we don’t go for tenders that specify an open source environment. It’s a one or other sort of thing, not a mixture.
So how open is open, and what’s the real cost?
Stewart is managing director of Wellington development house Intergen.