i noticed that e-text #16663 was "reposted" the other day. in .tei form. with a .pdf, an .html file, and a flock of text versions. so i took a look at it, and did some comparisons with the "original" versions, which were released earlier this month, on september 5th. i found the two .html versions looked quite a bit different. i preferred the look of the original, but that's personal taste. the second one -- by virtue of being the result of a conversion -- has the benefit of being "standardized", at least to some degree. (well, if we can assume that the conversion won't get changed.) the first one looked more like the .html had been hand-crafted, perhaps after an initial run through one of the .html converters. (though of course i can't say that with any degree of certainty.) if it's gonna be a practice to "deprecate" idiosyncratic .html versions that are being created by individuals over at distributed proofreaders, replacing them with versions that have been converted by machine, someone should start informing people not to make the original effort. when i brought up this very point a little while back, on the d.p. forums, juliet made a sharp public rebuke telling me not to "discourage" people. (i know, it's just not like her to do that in public, is it? but she did...) personally, i think _all_ the time people have spent over there making hand-tooled .html versions has been a big waste of energy; it would've been far smarter to perfect a text-to-html conversion routine instead... i see other instances, too, where distributed proofreaders appears to prefer to waste the time of volunteers to "keep 'em busy and happy", rather than ensuring that the workflow is as efficient as it could be... it must be nice to have that much energy at your command... one more thing: are we assured that the conversion routine is solid? or are "repostings" going to start happening with greater frequency? if you have to repost every separate variation from every .tei e-text every time the converter is changed, you could be doing that all day. for instance, in this version, there's a glitch on the chapter 18 header. when that's fixed -- which will only take a few seconds -- and the files are regenerated, which will also only take a couple seconds, will there be a whole other "reposting" then? because that will end up taking much more than a few seconds of time for more than a few people, including the people who mirror the library. i don't know the solution, other than to wonder why each of these variant files needs to be put in the library at all, if users will truly be able to generate them at will. which _is_ still the plan, isn't it? that they can generate them at will? setting all of their own options, in a simple way with a friendly interface? when do we get to see that part of the puzzle? and will marcello be telling people they can't have the font of their choice because p.g. "doesn't have a license to distribute it"? that would be a real bummer. this particular glitch is trivial from the viewpoint of usability of this text, but might be a sign that your quality-control checks need improvement. i caught this one because it had a black-on-white visible manifestation, but there could be many that don't, and they could come to haunt you. -bowerbird