The beauty of finished software

Relevant to complaints about Things and other software having not been updated in aaaaaaaaaages.


I think software can be finished from a features point of view (it is just a choice the developer makes after all), but it can never be finished from a maintenance point of view.

Like Marco has said on ATP these days we build our apps on quicksand – if you abandon it, it will simply stop working after a while as you don’t keep up with operating system, dependency, API, and other changes. Probably least the case in the iOS app store and most the case with web apps, but almost never zero anymore.

So if not for features, the currently ubiquitous subscription model is justified by that ongoing maintenance.

I am totally in support of stopping feature development of an app when it satisfies the initially set out requirement – things tend to get too bloated.


Feel the need to point out that Things 3 has had a LOT of recent updates, just hasn’t been “updated” in the sense of new features, fundamental UI design etc… and that’s probably good news.


The missing piece here is that an app can only be finished from the perspective of the developer and those that agree with their understanding of the domain that the app is for.

That’s simple in the case of the developer tools mentioned in the article.

It’s not for more complex domains.

I doubt, for instance, that GRR’s editors are happy with WordStar’s track changes features. Similarly I imagine that a lot needs to happen to the data in those WordStar files before it becomes the copies of Song of Ice and Fire sitting on bookshelves.

Thus, WordStar doesn’t actually cover the domain completely — it just happens to be finished covering the part that GRR cares about. And that’s good! But it is a narrow definition of “finished.”

Put differently: a tool is “finished” from the perspective of a user only when it has satisfied all of the user’s feature requests, and a tool can probably never be finished from the perspectives of all users.


Good points in this discussion, everyone—particularly that software requires constant maintenance to keep up with changing platforms and operating systems.

Same goes for roads, bridges, buildings, furniture… I love the idea of software, which we think of as being abstract and made of thought, as actually being part of the built world.

And @ryanjamurphy, you’re right about WordStar likely being only suitable for GRRM’s own needs and not the needs of the publishing production chain. However, given GRRM’s sales, I expect that the publishers are happy to have him use whatever software he wants. If he submitted his novels on punch cards, I expect the publishers would be cool with that. :slight_smile:

I have had friends who worked as science fiction editors, and I found their discussions of the business to be fascinating. Well into the 2000s, it was still cost-effective for writers to submit manuscripts as printouts, and then pay cheap typesetters working in India to just retype everything when it came time to turn the manuscript into a book. That was a less expensive and complicated alternative to dealing with the multiplicity of word processing formats available in the wild 20 years ago. (But Mitch, you say, even 20 years ago Microsoft Word was the standard. And yes it was—for business. But writers are often people who go a decade between software or computer upgrades.)


They’d be even cooler with it if his submission included the last two volumes of the seven-volume Song of Ice and Fire series, which he’s been promising since the fifth volume was published in 2011.

Maybe it’s time for him to move on from WordStar. Just sayin:wink: