isawnyu / awld-js Goto Github PK

Hi Seb,

When loading EOL resources, the EOL description has relative paths to image resources. When used on Open Context, clients look for the image resources on Open Context, not the EOL so as to avoid $)$ errors?

Perhaps modules/eol.js can be fixed (a hack) with something like:

description = description.replace("/assets/","http://eol.org/assets/");

Thanks!
-Eric

There is a problem while parsing RDF files with namespaces in tag names (eg dc:title). It seems this is due to a lack of namespaces support in jQuery (problem with colon escaping).
Maybe some tips here : http://stackoverflow.com/questions/853740/jquery-xml-parsing-with-namespaces
Thx for your help!

Currently, if the JS attempts to retrieve a resource from a supported domain, but doesn't get any content, it goes ahead and displays a blank popup on hover. It would be nice if "failed" links didn't exhibit this behavior.

github now serves everything over https. This is bad for showing off our current implementation because: #8 and #9

Many resources (including recently Pleiades) are moving to HTTPS, so people will be adding HTTPS URIs to new documents and wanting the script to work. Currently, it passes over HTTPS URIs e.g. for Pleiades entirely.

I was reviewing my own Mediawiki / AWLD bridge extension but I realized that 5 years without updates probably mean this library is not maintained anymore. Is this correct?

getting undefined and minimal data on pas resources, e.g. as linked from http://pelagios.dme.ait.ac.at/api/datasets/434fc5aea15f6d39040c1282c4a2df63/annotations?forPlace=http%3A%2F%2Fpleiades.stoa.org%2Fplaces%2F148217

Currently, if the page containing the JS and links is served over https but the individual links are coded with http URLs, only a blank popup displays on modern browsers (because of mixed content security practices). A better solution would seem to be: if the containing page has been served via HTTPS, rewrite any links to try with HTTPS; if 404 then don't display a pop-up at all.

In loading the test page, a lot of the Wikipedia resources randomly show or don't show based on the Wikipedia API endpoint returning HTTP 503 (Service unavailable) randomly when a lot of requests are made at once.

There are a few different approaches that might help fix this:

• set cache to true with requests (and make sure we're actually using cache?) so that we don't request all resources on every page load if they're already in the cache
• make a per-service request buffer and request resources serially to avoid overload
• only make requests on link mouseover instead of requesting all resources on page load (though this may result in unacceptable delay?)
• make a request error handler that can retry requests after a (random, exponential backoff?) delay