reprint count for medium and high

[CarolinOdebrecht]

We have different regulations on the reprint count:

The ODD says:

text has been republished very frequently since its original appearance

                                <valItem ident="medium">
                                    <desc>text has been republished occasionally since its
                                        original appearance</desc>
                                </valItem>
                                <valItem ident="low">
                                    <desc>text has not been reprinted since its original
                                        appearance</desc>
                                </valItem>

The sampling document says:

                <item>low: no reprints at all,</item>
                <item>medium: reprinted once,</item>
                <item>high: reprinted more than once,</item>
                <item>We will not include digitizations of texts in the reprint
                    count.</item>

This is crucial for "medium" and "high" The sampling document version is more restricted but clearer. It might be easier to just approximate reprint counts. We cannot assum that occasionally is the same for every language.

[CarolinOdebrecht]

This is linked with https://github.com/distantreading/WG1/issues/21

[lb42]

Certainly the ODD and the documentation should be in step! We don't seem to have yet reached any clear consensus on what counts as "high" or "low", and the numbers are likely to be different in different contexts anyway. We do however agree that we need those two values at least.

I propose to

• allow the value "unmarked", for texts where the encoder hasn't yet tried to assess the canonicity
• retain value "high" for texts " frequently reprinted during the period"
• retain value "low" for texts "rarely or never reprinted during the period"

At present the schema also allows for "medium", Should we keep that or change it to "unmarked" if used?

[CarolinOdebrecht]

A binary decision is easier to handle. Introducing a category "unmarked" is also a good idea.

[lb42]

So, at present, we allow high, low, unmarked, and unspecified. But if the value is unmarked it is ipso facto unspecified. And if after doing their best an encoder can only say something is unspecified, the effect for the user is just the same as if it was unmarked. The two are effectively synonymous. Since we use "unspecified" elsewhere, I propose to remove "unmarked" from the list of possible values and make the headChecker script convert any "medium" or "unmarked" values into "unspecified".

[CarolinOdebrecht]

Ok.

[lb42]

Closing this, as we are in agreement!