[tei-council] genetic draft -- from Brett, pt. 4
lou.burnard at retired.ox.ac.uk
Tue Aug 30 14:06:38 EDT 2011
On 29/08/11 21:22, Brett Barney wrote:
>> The @spanTo may be used to indicate the end
> Should read "The @spanTo attribute may be used to indicate the end"
> So<mod> can be either empty or not. Hmmm. Is there a reason for using a
> different approach here than with<add> and<del>? That is, why don't we
> two new elements,<mod> and<modSpan>? Alternatively, why don't we get rid
> <addSpan> and<delSpan> in favor of<add spanTo> and<del spanTo>?
I agree: we should deprecate xxxSpan.
> The camel-casing of<metaMark> is contrary to the rule implicit in other
> camel-cased element names, I think. To remember whether an element is
> camel-cased, I've always relied on a rule that goes something like "When
> stand-alone words (or their abbreviations) are combined, the second begins
> with a capital." As "meta" isn't a word on its own, shouldn't the element
> name be all lower-case? At the very least, the prose should be consistent;
> it's currently spelled "metamark," "meta-mark," and "metaMark" in various
OK, have standardised (I think) on "metamark".
> The prose description of metamark makes good use of a distinction between
> "text" and "document," but the<desc> of the element uses "text" for both
> senses. Worse, it says that metamarks can be "textual" or "graphical,"
> effectively cutting the pie yet another way.
"contains or describes any kind of graphic or written signal
within a document intended to signal how it should be read but not
forming part of the text itself"
>> Unlike marginal notes or other additions to the text, meta-marks are used
>> the writer to indicate a deliberate alteration of the writing itself,
> such as
>> ʻmove this passage over thereʼ.
> Two things: 1) I don't see how one can claim that "additions to the text"
> don't "indicate a deliberate alteration of the writing itself"; 2) plenty
> marginal notes say things like "move this passage over there," so I don't
> understand what distinction this bit is trying to shed light on.
Well, the idea is to distinguish a metamark from a marginal note. So are
you saying that they are indistinguishable? If so, why do we want
metamark? If you are not saying that, please propose some better text
for distinguishing them!
>> A metamark may contain text, or some other graphic which the encoder
>> to preserve,
> Should this be ". . . . which the encoder wishes to represent"? [I'm taking
> issue, I guess, with the implication that transcription and/or encoding
> preserve or destroy what's on the manuscript.]
"represent" is fine (though we do use "preserve" elsewhere with this
> I'm a *tiny* bit bothered by the lack of parallelism in the list of sample
> values for @function. Namely, "used" is not like the others.
Not sure what your point is here. Should it be a different attribute then?
>> The passage to which the metamark applies may be indicated in either of
> Without some sort of indication that the two methods cover different cases,
> don't like having two ways of doing the same thing. I can see that @spanTo
> couldn't accommodate teh example, but I'm not sure why @target couldn't
> for everything. If indeed both are needed for different cases, I think
> readers will appreciate something explicit about what the differences are.
@target can only work if it's possible to indicate the element/s
concerned -- so the passages being metamarked have to be discrete
elements. @spanTo works by just saying "everything from here to there".
> Further against @spanTo here, I'm uneasy that it means something different
> this element than it did on<mod>; here, it's not indicating the end of the
> current element but the end of the current element's referent.
@spanTo means exactly the same here as on <mod> -- I don't understand
why you think is different. What is "the current element's referent" ?
>> A further sentence was then added, while at some later stage the text and
>> also the metamark were deleted.
> I'm unsure whether it was added to the text or to the metmark. Upon close
> examination, I see that it's the former, in which case it dosen't seem
> pertinent information, except to explain bits that show in the image. But
> what about the multiple hash deletion marks or the other marginal tidbit
> one preceded by the cross mark)?
Agreed that this example is a bit too complex. Not sure how to cut it
> I was a little bit surprised to the<lb>s in the example's<ab>.
Me too. They're gone.
>> in section 1.4
> Should presumably read, for now, "in section ??"
>> Such cases should be distinguished from cases where the writer does not
>> intend to suppress the content
> I don't know why "should." I'm accustomed to statements that are less
> jugdmental and more context-aware: "If it is thought that blah blah blah,
> can etc. etc., but if instead it is thought that yakkity yak, one can this
> that." >
>> but these should be distinguished from the ʻdeletionʼ signalled by the
> 2 things: 1) Here and in the next two paragraphs "cross" is used where I'd
> have expected something like "X." Just checking that you really do mean to
> say "cross." 2) If the whole point rests on the notion that this is
> from a deletion, I hope there's a way to avoid describing it as one, even
> surrounded by quote marks. What about something like "but these are
> semantically different from the large cross/X/intersecting basically
> lines, . . . "?
OK, have reworded discussion of this example a bit.
>> instead the status of existing material changes
> I think I see the meaning here, but I suspect that many will not. It might
> easily be argued, in fact, that retracing is done precisely to *not* change
> the status of the existing material.
ok: changed "changes" to "is reaffirmed"
>> which is subsequently fixed,
> At least where I am, "fixed" is seldom used in the sense intended here, and
> the most common sense of the word is unfortunately almost opposite what's
> intended. What about "confirmed" or something similar? If pressed I could
> even make the case that "fixed" isn't really the right word anyway, since
> can imagine cases in which the second act of inscription uses a more
> ephemeral (but possibly brighter or bolder, etc.) medium.
"fixed" -> "reaffirmed" (for consistency)
>> its @cause attribute may be used to distinguish them.
> I have an ominous feeling that I may be unwittingly stepping into a
> minefield, but here goes. Why not @reason?
Good catch. I dunno. @reason is used (explicitly) on <gap>, <supplied>,
<surplus> and <unclear>. @cause is used on <lem> <rdg> and <rdgGrp>
>> we can distinguish at least two attempts
> I think this should read "we can distinguish at least *three* attempts"
>> When metamarks and other markup-like strokes are inked over with the same
>> purpose as the fixation or clarification of text passages, the redo
>> described in the next section should be used in preference.
> I'm not following this. In any case, it seems to contradict the definition
> given for<redo>: "points to a marked-up intervention in a text which has
> subsequently been marked for a second time *in a different way.*" Wait. I
> think I get it, now that I look at the Faust example. "in a different way"
> definitely misleading, I think. And so is "with the same purpose as the
> fixation or clarification of text passages." How about just "When metamarks
> are retraced, the redo element described in the next section should be used
> in preference"?
You don't object to that "should" :-) ? Anyway, have deleted the
>> The element restore (??)
> Not sure what the question marks here signify. Is the idea that maybe we
> don't want/need<restore> at all?
No, it's a formatting artefact. There's a cross ref to the place in the
existing PH where <restore> is described.
>> <line>This is<del layer="#s2" rend="overstrike"> <undo spanTo="#Xa"
>> rend="dotted" layer="#s3"/>just some<anchor xml:id="Xa"/> sample<undo
>> spanTo="#Xb" rend="dotted" layer="#s3"/>text,<anchor xml:id="Xb"/> we
>> need</del> <add layer="#s2">not</add> a real example.</line>
> Hmmm. OK, I know that this isn't the example you want to go to press with,
> but it seems to make problematic the way<undo> and its attributes are
> described a little earlier. Specifically, the<undo>s here don't actually
> "point to the marked-up intervention," at least not in any way that the
> I was expecting from the definitions. In the example, there aren't any
> @target attributes at all; it would seem that the logic for knowing what's
> being undone depends on parent/child relationships, right? But that isn't
> ever explained. A bigger concern, which I'm just not smart enough right now
> to try to answer, is whether that's even a water-tight system. Will it
> be the parent that "names" the intervention?
This example uses @spanTo to indicate the scope of the undoing, so I
agree that the "point to" language in the <desc> for <undo> is wrong and
have changed it. You are right that we need to explain better what is
being undone in this case: I think the intended rule is (as you surmise)
that it's the parent element, so I've added the following to the example:
"Note that in this case, since <att>target</att>
is not supplied, it is the effect of the parent element (the
<gi>del</gi>) which is assumed to be undone. "
More information about the tei-council