Backwards Compatibility v. Support for the Latest Features

One of the dilemmas facing all software developers at one stage or another, is whether they should support the latest operating system capabilities or whether they should forgo those enhancements for the sake of backwards compatibility.

So far it has always been my policy to support the latest Mac OS X versions right from their release date, but always maintain backwards compatibility with the previous operating system version. At this point in time, this translates to all my products supporting both Mac OS X 10.3 and 10.4.

Apple’s release in quick succession of 4 major OS revisions has brought us a host of new features, but also a real upgrading headache. The vast majority of current publicspace.net customers seem to quickly upgrade to the latest versions of the OS and pretty much everybody seems to be on Mac OS X 10.4.5 right now.

Or at least, that’s what it looks like from here. In reality, it is difficult to gauge exactly what percentage of users have migrated to the latest version and how many people are lagging one, two or even three major system upgrades behind.

Precise information about how many Mac OS X users are still using Mac OS X 10.3 is hard to come by. OmniGroup are publishing the OS version data collected on a strict opt-in basis by their Software Update online service and this suggests that 94.5% of all Mac users are running some version or other of Mac OS X 10.4. Of these almost 40% have the latest point update and almost 70% one of the two latest point updates.

This data is, however, perhaps not really representative, as it only includes OmniGroup customers that have decided to opt into this scheme and are happy to download the latest updates via the internet. Are these people particularly tech-savvy? Are they the ones that have the fastest internet connections?

It seems certain that this type of automatic update service will be used substantially more by those who always want the latest version of everything and that the data will be significantly biased against those who update infrequently.

Right now, there are several features of Mac OS X 10.4 that would allow me to build “a better” version of both A Better Finder Rename and A Better Finder Attributes. Moreover, testing that each release still works correctly on Mac OS X 10.3 is time-consuming and that time would perhaps be better spent on adding new features and enhancing existing ones.

This is why, I am very tempted to discontinue support for Mac OS X 10.3 for the next versions of A Better Finder Rename and A Better Finder Attributes. Surely users who have not yet upgraded to 10.4 can still download and use the 7.2.1 or 4.0.2 versions of the tools? But is that really good enough?

Why don’t you help me decide by voting in the attached poll, adding a comment to this post or contacting me privately via email?

2 thoughts on “Backwards Compatibility v. Support for the Latest Features

  1. My suggested rule-of-thumb, and I am by no means a developer, is support for the latest and previous versions of any OS; in this case, 10.4 and 10.3. For those of us in the corporate environment, OS updates are not made every 12-18 months as is Apple’s release schedule. Therefore, a reluctance is built towards purchasing software which “changes” so much as to make its support and compatibility an issue in a short time frame.

    Just my thoughts.

    P.S. – When is version 8 going to be released?

  2. Thanks, that’s exactly the rule of thumb that what I’ve been following until now.

    Version 8 is probably still at least 18-24 months away. I don’t want people to have to pay for upgrades all the time, so I am planning on releasing a “half point” update every 4-6 weeks, e.g. 7.2, 7.2.5, 7.3, 7.3.5, etc. That should keep me going for a while.

    Corporations are usually slow upgraders with many still using NT 4 a decade after its release..

Comments are closed.