I'm not sure we really do disagree, so much as I didn't explain my point well enough. I agree completely with what you say - the maintenance overhead is one di

I'm not sure we really do disagree, so much as I didn't explain my point well enough. I agree completely with what you say - the maintenance overhead is one disadvantage of the many specialised tools approach. But, my point is, assuming this is a very rare activity, that maintenance cost is paid infrequently and is therefore not, cumulatively, very high. If you ski every day however, carrying one tool for each feature indeed starts to bring too much maintenance overhead and you'd at that point opt for more all-in-one convenience. But the user who does a specialised activity as the norm is very rare, thus such products don't really have a market. I wouldn't necessarily describe this 'many specilised tools selected individually for a known specialised use case' approach as being in line with the UNIX philosophy, which is more 'many specialised tools with the ability to be chained together to handle any use case'. In my example of skiing, the tools don't need to be chained. They can, and preferably should, be quite isolated in their function and use. They're all completely vertically functional - if one goes down, the others all still work fine and only one feature is lost. It's just about unbundling the completely orthogonal features you want and picking options tuned to that specific use case, sacrificing some convenience for better quality in only that use case. The all-in-one iPhone type solution is more about bundling regularly-used (though not always used, in each case) features to work across a broad range of more general use cases, necessarily sacrificing optimal quality in any particular use case for convenience across all of them.