I think Would Not has got it right. I too am an engineer, but in the end it's mostly a matter of working methods.
Engineering works to very close tolerances (a) because it's frequently a functional requirement, and (b) as he says to enable interchangeability.
But unless you're on first fix carpentry there's not too many woodworking joints that will look or perform decently with up to 1/32 in marking and making variability on each part, especially not if (to speak engineering) the tolerances on the two mating parts add and you get a gap of 1/16in.
The reality is as somebody else has said that most joints are actually fitted to within a thou or so. The first half of the joint e.g. the mortice may well be made to quite an approximate tolerance, but the mating part will be hand fitted using a shoulder plane or whatever to match.
This probably won't deliver interchangeability, because even though the fit is very finely adjusted the dimensions will probably be different. Not to mention that the completed item will probably subsequently move all over the place with changes in humidity.
This hand fitting is in engineering/manufacturing known as craft production. (pre-Henry Ford) If on the other hand you are engaged in repetitive manufacturing - even of wooden items - then it becomes a matter of economics that parts are interchangeable. Nobody is going to be able to afford to spend loads of time on each item either hand fitting, or selectively matching parts that happen to fit.
Does it matter? Personally i'm very happy to be able to set a cut to within a thou using e.g. an Incra positioner, and then make the cut on a machine that's accurate and rigid enough that it's to dimension and square. More or less that is, because at our level of wood machines are still pretty approximate by engineering standards.
There's a huge level of skill involved in working around the vagaries of inaccurate machines - this starts at the design stage, and if you have to adjust/fit parts then sequencing is important too.
But is it necessary to do this to be credible? I don't think so because even so called accurate kit needs very careful set up and use if you are to even hope to get parts off it that are even close to simply slotting together without fitting. We're quite some way off machines that can do this without an intimate understanding of what's going on. People seem to demand ever higher standards of fit and finish anyway, and it's just another skill set...
It's not of course a 'good thing' in itself i think either - there's not much to be had from obsessing about this one aspect of making when it's possible to produce stuff that has it's own unique look and feel, and is fully functional despite much less precision. Nor is there anything that makes it great (art?) just because you spend hours slaving away over it, other wise we'd all be working with nail files!
Maybe it's just a matter of personal preference, of what you are trying to achieve (how you mix art and economics) - with the possibility of either approach (within reason) producing great work???
Engineering works to very close tolerances (a) because it's frequently a functional requirement, and (b) as he says to enable interchangeability.
But unless you're on first fix carpentry there's not too many woodworking joints that will look or perform decently with up to 1/32 in marking and making variability on each part, especially not if (to speak engineering) the tolerances on the two mating parts add and you get a gap of 1/16in.
The reality is as somebody else has said that most joints are actually fitted to within a thou or so. The first half of the joint e.g. the mortice may well be made to quite an approximate tolerance, but the mating part will be hand fitted using a shoulder plane or whatever to match.
This probably won't deliver interchangeability, because even though the fit is very finely adjusted the dimensions will probably be different. Not to mention that the completed item will probably subsequently move all over the place with changes in humidity.
This hand fitting is in engineering/manufacturing known as craft production. (pre-Henry Ford) If on the other hand you are engaged in repetitive manufacturing - even of wooden items - then it becomes a matter of economics that parts are interchangeable. Nobody is going to be able to afford to spend loads of time on each item either hand fitting, or selectively matching parts that happen to fit.
Does it matter? Personally i'm very happy to be able to set a cut to within a thou using e.g. an Incra positioner, and then make the cut on a machine that's accurate and rigid enough that it's to dimension and square. More or less that is, because at our level of wood machines are still pretty approximate by engineering standards.
There's a huge level of skill involved in working around the vagaries of inaccurate machines - this starts at the design stage, and if you have to adjust/fit parts then sequencing is important too.
But is it necessary to do this to be credible? I don't think so because even so called accurate kit needs very careful set up and use if you are to even hope to get parts off it that are even close to simply slotting together without fitting. We're quite some way off machines that can do this without an intimate understanding of what's going on. People seem to demand ever higher standards of fit and finish anyway, and it's just another skill set...
It's not of course a 'good thing' in itself i think either - there's not much to be had from obsessing about this one aspect of making when it's possible to produce stuff that has it's own unique look and feel, and is fully functional despite much less precision. Nor is there anything that makes it great (art?) just because you spend hours slaving away over it, other wise we'd all be working with nail files!
Maybe it's just a matter of personal preference, of what you are trying to achieve (how you mix art and economics) - with the possibility of either approach (within reason) producing great work???
