[tei-council] Feedback on genetic editions draft

[Stuart Yeates]

In giving this feedback I'm very aware that the people who seem to me be most enthusiastic about this proposal appear to be from the humanities side of the digital humanities community and most of the people who seem the most sceptical (including myself) seem to be from the sciences side. I'm not sure whether this is necessarily symptomatic of anything deeper than groups talking past each other (a la The Two Cultures https://secure.wikimedia.org/wikipedia/en/wiki/The_Two_Cultures ), but it worries me, none the less. Some of these points are as much reflections designed to ease communication, rather than action points.

* The name 'genetic editing' is very confusing to me. I already have three completely separate kinds of genetics in my head (uni-cellular, multi-cellular and algorithmic) each of which use the same terminology, I know the linguists have another. Writing https://secure.wikimedia.org/wikipedia/en/wiki/Genetic_editing has helped me understand, but anyone from a sciences / programming background is likely to have similar issues. Avoiding defining 'genetic editing' in the draft because there are many definitions forces people to research this for themselves, which is not helpful.

* The draft seems to have many examples going from physical manuscript TEI encoding, but very little from TEI encoding to some system users interact with. The example in 4.2 Genetic Graphs is start, but we need more.

* There is a lack of motivating examples. Even references to other projects with which have done cool stuff in this space would be good. Even unfunded project proposals that describe a potential system.

* The diagram in 4.2 Genetic Graphs is crying out for an arrow indicating the passage of time, if I read it correctly. Does time flow from left to right in this figure?

* Use of "http://edition.net" in URLs is wrong on so many levels. This is the exact use-case for example.edu and friends. See http://en.wikipedia.org/wiki/Example.com

* I have the sense that this is really a case for standoff markup in some form rather than making life much more complex for everyone who doesn't care about genetic editions. The current proposal appears to make life much more difficult for people who aim to provide sane handling of a broad range of TEI documents, by substantially increasing the complexity of tags, attributes and cross-references needed to be handled. Is there a plan to implement a stylesheet to derive an old-style TEI document from a new-style TEI document?

* It is disappointing that  rend="underline" is still needed. Isn't the point of genetic editing is that we now have the tools to express the location of the underlining relative to the text in a formal way which obligates the need for such hacks? (We still consider @rend a hack, right?) Or is there a granularity below which we switch from genetic editing back to normal encoding? Where do we describe that level?

* Many of the features of genetic editing seem to be very challenging to provide to the user in accessible manner (for example present to blind user). Many institutions (such as my own) and sources of funding (including some which we have accessed) have legal responsibilities around accessibility, which would limit our use of these features in some projects.

cheers
stuart