These videos of some very cool solo viola studies composed and played by Garth Knox are vivid illustrations of what a virtuoso performer with some serious compositional chops can do. I look forward to hearing more from Knox in forms beyond the technical etude.
The rich diversity of composed music is due in part to the wide range of experience and imagination that each individual composer is able to bring to writing for particular instruments, voices or ensembles. Of those composers who write particularly well for instruments (let's be honest: not all do), many are generalists, with broad knowledge of each of the instruments they compose for, while others are themselves are more specialized — often as both performers and composers — on particular instruments. I don't think either generalists or specialists have an edge here, but rather that our musical lives are livelier because we have both. We benefit from the virtuoso pianist- or violinist- or percussionist-composers, but also from someone like Berlioz, whose own instrumental skills (he played flute, flageolet, guitar, and timpani) were not necessarily prerequisite or even terribly relevant to his own virtuosic writing for orchestra. Likewise, we benefit from having both a Mahler — whose scores were micro-managed with technical details — and a Sibelius — who managed to create large works of comparable detail and complexity with an efficient minimum of notation.
For the generalists, not knowing many specifics about the techniques which may be brought into play in order for players to realize their scores, can often be an efficient and profitably open situation, an invitation for players to bring their own experience more closely into the rehearsal process. Some generalists make a point of not knowing too much about instruments, on the principle that knowing too much usually means knowing overly restrictive limitations or settling into stereotypical writing for an instrument. The obvious risk of not knowing too much is not knowing enough and then writing something that is impossible or simply dull for an instrument. On the other hand, some generalists are practically encylopedists when it comes to technical matters, adding very fine details and specifications in their scores. As admirable as this is, it can also carry a bit of a risk. For example, a fingering for a particular tone may work on one mark and model of an instrument but not on another. For myself, although I'm perfectly capable of adding complete bowings, fingerings, breath marks and more to my scores, I hope that I am able to restrain myself somewhat in recognition of the limitations in my knowledge. My bowings or fingerings may well "work" reliably enough, but I gladly differ to the expertise of players whose experience and imagination can be drawn upon to often offer better solutions.
And then there are the specialists. A composition grounded in a performer/composer's own virtuoso playing technique is a legitimate and honorable genre. (I am a particular afficionado of the earliest virtuoso literature, for recorder, cornetto, viol, and virginal.) In many cases, the particular mechanics of a given instrument may be such that that the generalist may be wise to defer to the specialists altogether. For example, writing for the accordion — in any one of its daunting number of variant forms — means either writing in a very general, even basic, way for the instrument (which I have done), or making a commitment to working very closely with a particular player and a particular instrument (which I have not yet done (never say never.)) Likewise, I am very hesitant to write specific woodwind multiphonics, unless the project is writing for a particular player on a particular instrument. I might also consider using a notation which was open to a variety of multiphonic solutions within some specified characteristics.