By a lucky coincidence, these URLs still work: the & at the start of the entity still separates the ID from the rest of the URL parameters, and the dl.acm.org endpoint must ignore invalid parameters (in this case, the amp;type=pdf), e.g. even http://dl.acm.org/ft_gateway.cfm?id=1454225 works fine to download the paper.
When adding a paper from dl.acm.org, such as https://dl.acm.org/citation.cfm?id=1454225
the download URL is automatically filled out to point to https://dl.acm.org/ft_gateway.cfm?id=1454225&type=pdf . However, the URL undergoes HTML escaping (that is, escaping for inclusion in a HTML document, not even escaping for inclusion in a URL), so the
&
gets converted to the HTML entity representing the ampersand:&
: http://dl.acm.org/ft_gateway.cfm?id=1454225&type=pdfBy a lucky coincidence, these URLs still work: the
&
at the start of the entity still separates the ID from the rest of the URL parameters, and the dl.acm.org endpoint must ignore invalid parameters (in this case, theamp;type=pdf
), e.g. even http://dl.acm.org/ft_gateway.cfm?id=1454225 works fine to download the paper.The database likely has many of these corrupt download links, e.g. http://gnosis.stellargraph.xyz/catalog/paper/1241/ is the stored copy of the above paper, and the download button is:
(Note the second
&
HTML escape, too.)