irregular use of distributions

[EnnoMeijers]

The Globalise project publishes a dataset (https://datasetregister.netwerkdigitaalerfgoed.nl/show.php?lang=nl&uri=https%3A%2F%2Fhdl.handle.net%2F10622%2FLVXSBW) that contains a very large amoint of distributions (>6800?). This seems unbalanced with the regular use of distributions and might lead to problems for applications reusing the datasetregister data. It feels like the dataset is described on a too granular level. Is this type of use aligned with the intentions of the DCAT specification or our intented use?

[coret]

@LvanWissen I think the transcription dataset is better usable if it is provided in one compressed file?

[LvanWissen]

But that's not how Dataverse publishes the data (and can ingest data, there is a file size limit).

[coret]

@LvanWissen https://support.dataverse.nl/support/solutions/articles/80001022346-upload-of-large-files-a-lot-of-files-in-dataversenl states the "double zip" option, it this a possibility for the transcriptions?

[LvanWissen]

Right now, it's very easy for a user to download the PageXML of a single inventory number. With that goal in mind, we already double zipped this data so at least the PageXML for a single inventory number is packed together.

We could have opted for a grouping until the maximum upload size of this Dataverse would have been reached: 2GB, but that would have hindered the usability. And this way you still end up with a couple of files/distributions.

All possible, but for what goal are we making these dataset descriptions? A machine is now perfectly capable of downloading the files, while a human can easily pick a single inventory number. Can we make guidelines or a best practice?