jhu-digital-manuscripts / AnIOp

to track the activities of the Mellon funded Annotation Interoperability project
0 stars 0 forks source link

Some text alignment issues #102

Open jabrah opened 4 years ago

jabrah commented 4 years ago

I wanted to record some issues that Chiara has noticed with some of the annotations. Some may be our client side rendering, some may be the annotations themselves. To be investigated.

jabrah commented 4 years ago

Regarding point 2 above, This annotation seems to have https://aniop-atlas-staging.eldarion.com/wa/urn:cite2:hmt:msA.v1:34r/translation-alignment/605/ a CTS URN embedded in the first annotation body

jacobwegner commented 4 years ago

@jabrah inline responses below:

f. 33v, the manuscript starts with v. 2.463), but the transcription and translation start with v. 2.459

The translation alignments are organized by English sentences, so there isn't always a 1:1 map to the line numbers of the greek transcription.

So in this case, the English sentence:

2.459-2.465) And as the many tribes of winged fowl, wild geese or cranes or long-necked swans on the Asian mead by the streams of Caystrius, fly this way and that, glorying in their strength of wing, and with loud cries settle ever onwards, 1 and the mead resoundeth; even so their many tribes poured forth from ships and huts into the plain of Scamander,

is mapped to lines 2.459 to 2.465 in the Greek.

We could provide a separate transcription annotation that would start at 2.463 if desired, but the sentence alignment will always be organized around the English sentences

jabrah commented 4 years ago

Ah, a misunderstanding on my part. I think I get it. I feel like this ties into a broader discussion about the usage of these annotations in our client and isn't something that should be quick-fixed just for user testing.

jacobwegner commented 4 years ago

@jabrah wrote:

F. 34r, a cts urn shows up in the text with no reason (actually, I think there is a reason, because the ms. text differs significantly, but it just looks odd)

The source of the data is from https://docs.google.com/spreadsheets/d/1-Zz1TCm1bVygngmWeXuskzlfQHN3ER6HERX0DC-VRI0/edit#gid=283848305.

It looks like that there is an update to the data that fixes the issue:

image

I'll work to get our copy of the data updated, which should fix that one.

jacobwegner commented 4 years ago

@jabrah wrote:

F. 34r, the ms. ends at v. 2.512), the text at v. 2.515). This last issue is the most relevant, because it breaks the alignment between ms. and text in the following pages, making it impossible to read both up until f. 37r

Yep; another "side-effect" of the English sentences being mapped to Greek lines (which are in turn mapped back to the folios.

jacobwegner commented 4 years ago

sigh...I misspoke; those errors are in the original spreadsheet (as seen in my screenshot); I've just gone ahead and fixed them in our copy and redeployed

jabrah commented 4 years ago

Parking for now, to be revisited in the near future (i.e. after user testing)