'Add new work' issue

[MandyShaw]

Happy New Year all.

I've just added a 12 part setting of In te Domine speravi by Hassler, which demonstrated a bit of a problem with the 'Add new work' functionality.

There was, in fact, a 4 part setting with same title & same composer on CPDL already.

'Add new work' attempted to merge the new with the old, which was obviously incorrect. I was only able to resolve the problem by removing the changes that had been automatically added to the old page; moving the old page appropriately; editing the composer page appropriately; and creating a new page for my new score. I could create the new page easily enough because I always provide my email address on the 'Add work' form and I therefore had a readily available copy of the necessary wiki text.

However I cannot help thinking that a newish CPDL contributor could have got in a major tangle over this, especially if they had not specified their email address on the 'Add work' form.

I'm not claiming to have thought of a solution to this, but would it be possible, as a start, to check that the number of voices matches before merging two works?

Meanwhile, many thanks for adding the Lyricist.
Mandy

[choralia]

Happy new year to you, too!

Unfortunately we also have many cases where the same work includes editions with different voicings that are supposed to stay on the same work page instead. I'm not sure that different cases like these can be treated automatically. Many other similar situations exist, not only related to voicing, that are very hard to recognize and handle with automatic means. This is the reason why the category Pages with recent additions was created: newish CPDL contributors can leave things as they are, and more expert users (not necessarily administrators only) can check this category from time to time and fix pages yet to be finalized. The possibility for any users (and not only for administrators) to finalize submissions by less expert users is one of the key features of the new "automated" procedure, aimed at distributing workload and reducing submission backlog.

Max

[MandyShaw]

Understood. I do wonder, though, whether the process could put up a confirmation prompt before merging, allowing the user to do a quick sanity check? Just a thought. Alternatively I would suggest perhaps the merge activity might automatically open a window with the wiki text in, just in case it is needed to clean up & the user has not entered their email address.
P.S. if I were a new user and had left things in a mess on a page created previously, even if someone would come along eventually & clean up, it would worry me a lot - this is different from the previous setup where the edition would not be visible /at all/.

[choralia]

just in case it is needed to clean up & the user has not entered their email address.

The user must always enter his/her email address: the submission form is not accepted otherwise. So this case shouldn't occur.

this is different from the previous setup where the edition would not be visible /at all/.

On the other hand, the previous setup disappointed users who were willing to see their editions published immediately, while they had to wait quite a long time (even several weeks) due to the backlog that administrators were unable to clear. As both procedures remain available, users are free to choose the one that they prefer.

Max

[MandyShaw]

I know, the new approach is much, much better overall, I am really just concerned that the merging happens with no warning. I know one should check in advance for existing editions, and that my failure to do so was the root cause of the trouble I had, but I do still think some sort of confirmation prompt would be good if an existing edition is detected, so that the user could cancel out of the whole thing if concerned (& then use the old approach if preferable).

[choralia]

I think that adding a warning/confirmation step may scare users when they do something that is totally normal, i.e., when they add a new edition to an existing work page. Being a very frequent case, this may cause a significant number of incomplete/aborted submissions, more difficult to identify, recover, and fix. I would expect much fewer cases instead where the user, by mistake, submits a new work using precisely the same name of an existing (different) work. I'd prefer to treat the latter cases as occasional exceptions to be fixed manually, rather than adding complexity and possible shortcomings to one of the most normal and frequent submission cases.

Max

[MandyShaw]

I bow to your greater knowledge of the statistics - anyway I will take more care in future.

[choralia]

This is just my opinion. Let's see what others think. I definitely prefer a wrong submission that can be fixed, rather than an aborted submission that leaves no traces and therefore cannot be even identified.

Max

[vaarky]

How about putting up something after the thing has been done, asking the user something like the following?

"Does this look right or is anything obviously wrong? Please describe here and your answer to this form will be e-mailed to notify CPDL.org administrators."

That could prevent submissions lost due to user uncertainty, but would give the contributor a way after the fact to raise the alarm with the addscore crew with hopefully helpful detail, helping the people who are not sure how to contact CPDL admins.

[choralia]

I think that the current approach is equivalent to the one that you're proposing, Vaarky. Whenever a submission is made, the new or amended work entry is automatically tagged as "to be reviewed by an expert user". This is done using the "Pages with recent additions" category template, which is automatically added to the composer page, close to the the new or amended work entry. Then, if the contributor is expert enough to autonomously review and finalize the submission, he/she can also remove the "Pages with recent additions" category template, and the process is complete. If the contributor is not expert enough, the work entry remains tagged so that an expert user (not necessarily an administrator) can review and complete the submission. This would be also valid in the case the contributor notices that there is something wrong: just do nothing, and the submission remains tagged so that an expert user will soon review and finalize it.

At the moment, only the work entry is tagged as "to be reviewed by an expert user". The corresponding work page is not tagged as it appears redundant to me, since the work page can be immediately accessed from the work entry. We can possibly add the "to be reviewed by an expert user" tag to both pages.

The obvious assumption behing this process is that administrators and expert users check on the "Pages with recent additions" category from time to time, and fix incomplete/wrong submissions. Thsi worked so far: the backlog of submissions to be reviewed is always very short. At this very time, there is only one pending submission in the list. If we can clear the backlog of submissions made with the old process, the backlog problem is eventually defeated.

Max

[vaarky]

Thanks for the extra context and for all your work on this. Sounds good to me!