Feedback from TDWG conference

[PaulKiddle]

Feedback from the TDWG 2020 conference included:

• Whatever new solution you choose, I would take into account as requirement how easy it would be to migrate again in the future
• Other platforms to consider
  • PlutoF (About | Repository - last update 2016?)
  • exploration include the ALA’s Profiles module ? It currently supports Flora of Australia which is more scratchpad-esque
  • DINA - there is a more up-to-date software than DINA Collections (See gitub org - not sure which repository specifically)
  • Symbiota
• Lots is possible with static website generators (such as Jekyll) too, including expressing data (csv, json, yaml) as pages
• Desire to keep backwards compatibility to DWCA format
• Compatibility with DOIP and FAIR DOs would be nice

[mjy]

@PaulKiddle just got to seeing your talk on youtube tonite. Thanks for the summary. Some random thoughts:

• Have you tried simulating random data on any of the target platforms, specifically to look at scaling issues? You mentioned that you wanted to move to a Service + Multi-Tenant architecture rather than multiple instances, this might have scaling considerations?
• Have you identified a target breakdown in the quantity of effort you want do spend, as developers on front. vs. backend if things were "optimal" in the future? For example 20% back-end, 80% front end?
• Have you considered targeting multiple platforms to get the data out from the current SPs, then focus on developing display-only platforms that play against a suite of APIs, perhaps adding social element (Discussion, email, feedback) etc. but not handling data curation tools? So farm the data out to whatever platform best fits, then work to aggregate data to fixed, "static", high performance sites.
• Are there current curators that would be OK with jus curating data and serving it by API/Export mechanisms, without "fancy" public facing websites?

[PaulKiddle]

Hi Matt,

Thanks for these thoughts and questions, I'll do my best to answer them:

• Just before the conference I had started to import real data from one of our sites to check how easy it would be for a single site. Once that's done I'd like to start testing with multiple sites to see how it scales. I've been basing my thoughts on the idea that running multiple services would scale worse than a single multi-tenant service, though you might have a better idea of whether that's true?
• We haven't identified a target development breakdown. I don't think there is any definite line we can draw here to be honest, though as a rough figure 80/20 would be better than our current Drupal situation.
• I think using multiple different platforms for different kinds of data/service is a good idea. The team have had concerns that a multiple platform model might be too much overhead, but as long as we don't have too many I don't think it will be a problem - and in fact will probably be necessary.
• Static sites and export-only might be an option, though many users like Scratchpads for its data management/editing interface. We're planning to survey our users to get a better idea of exactly what they need/want from Scratchpads.

As for the API, I haven't yet had a look at those new endpoints, thanks for the update. So far my testing has been either through the Ruby CLI interface, or using a very basic RPC endpoint I added. Once I've got a better idea of how we'd like to importing and query the data I can get back to you about those other endpoints.