Open jonathanrobie opened 7 years ago
I wonder if "maintain sentence order" can be formally specified as: "if all Treedown labels and annotations are removed, one is left with the original words of the text, in their original order" or similar.
I wonder if "maintain sentence order" can be formally specified as: "if all Treedown labels and annotations are removed, one is left with the original words of the text, in their original order" or similar.
That's a good way to explain it.
If the goal is "Easy to exchange in email, issues lists, and forums," then we also need to consider text input on mobile devices. In particular, this may have some bearing on the tabs vs. spaces issue.
Some comments from today's meeting:
James Tauber One thing I’d like to clarify is that “Based on a linguistic model easy to explain to students learning from traditional grammars” doesn’t mean we resort to traditional analyses and ignore a century of descriptive linguistic work but just that we use broadly agreed upon descriptive terms, avoid framework-specific jargon, etc
Jonathan Robie Ideally, I would like it to be usable by someone who (1) learned via Smyth / Robertson, (2) learned via Mounce, (3) learned via Porter, (4) learned via Runge / Fresch/ Black / Buth
James Tauber it’s one reason I’ve been looking stuff up in CGEL, NOT because an analysis of English is necessary relevant to an analysis of Greek but because they tend to be modern and precise without being jargon-heavy or framework-specific
Christopher Land The main categories are fairly generic at this point, even if different models view them differently or label them differently
Micheal Palmer And where we need to refine a term, we can do that as long as we document what we mean and make it available to all.
We discussed granularity of representation with respect to goals, and agreed that the ability to represent the same text at different levels of granularity is an explicit goal for both doing the analysis and consuming it. A person who is marking down a text might start coarse, then add more detail in successive passes. A person who is reading a text that is fully marked down in the system might start by seeing a coarse breakdown, then drill down to specific parts of the sentence for more detail.
For instance, here is a coarse representation:
s πολλοὶ
v ἐπεχείρησαν
od ἀνατάξασθαι διήγησιν περὶ τῶν πεπληροφορημένων ἐν ἡμῖν πραγμάτων,
The same text in the next level of analysis:
s πολλοὶ
v ἐπεχείρησαν
od
v.inf ἀνατάξασθαι
od διήγησιν
+ περὶ τῶν πεπληροφορημένων ἐν ἡμῖν πραγμάτων,
And in the next level of analysis:
s πολλοὶ
v ἐπεχείρησαν
od
v.inf ἀνατάξασθαι
od διήγησιν
+ περὶ
τῶν
+ v πεπληροφορημένων (: Embedded clause - using adjunct :)
+ ἐν ἡμῖν
πραγμάτων,
One goal, not explicitly mentioned here but I think mentioned in conversation, is ability to embed treedown in markdown documents (and have markdown formatters that support extensions be able to do something custom with that treedown)
One goal, not explicitly mentioned here but I think mentioned in conversation, is ability to embed treedown in markdown documents (and have markdown formatters that support extensions be able to do something custom with that treedown)
Yes, I have been doing that in some of the examples I have been sharing, and I think it is important.
We should have a high level goals section. Here's my first shot - let's try to come up with a list we agree on.
Expressivity
Usability