Closed sbmlsecretary closed 2 years ago
Original comment by: luciansmith
Thanks! This is now fixed in SVN for L3v2, and will be incorporated into the forthcoming specification.
I'm not sure what the best thing to do for L2v5 is. Would it be possible to make this change, and incorporate it into a new release of the spec, instead of a whole new version?
Original comment by: luciansmith
Yes, I believe it would be acceptable to issue a new release of the specification. The SBML Development Process says this about releases, and making a new release for this issue is perfectly in line with this principle:
"After a period of time, an accumulation of issues may warrant the development of a new release of Version 4, ..."
(and yes, it needs to be updated to talk about version 5.)
Original comment by: mhucka
Original comment by: luciansmith
With the release of L3v2, this issue is fixed there. However, it remains in l2v5, so I added a tag for that specification instead, so we can add it to the list of errata there, possibly for a later release of that spec.
Original comment by: luciansmith
This is now fixed in the SVN for L2v5 as well; whenever we do a new release of that, those changes will go in.
Original comment by: luciansmith
Please update the identifiers urls to the https versions. Unsecure connections should be avoided whenever possible (and could also become unavailable at some point, e.g., as happened when EBI moved completely to https).
Otherwise agree with the change.
Original comment by: matthiaskoenig
I agree with Matthias - change to https and then okay
Original comment by: sarahkeating
I agree with this change
Original comment by: chaouiya
There are a lot of 'http://' lines in the SBML spec, and I'm not sure how many can and/or should be changed to 'https://'. Can we get away with changing all of them except the namespace URLs?
Original comment by: luciansmith
I agree with the change. Regarding http://identifiers.org or https://identifiers.org in namespace urls, I would not consider it is a major issue at least for now. Both ways would be ok. The identifiers.org website now redirects http to https and this redirection service should be a feature maintained by identifier.org for a long term.
Original comment by: zhangfen
OK, went through all the URLs in the spec and updated them to use https if possible. This was, notably, not possible with sbml.org and biomodels.net and a few others; identifiers.org was one of the few ones where https actually worked. However, that was for l3 and l2; adding a new issue for that.
Original comment by: luciansmith
Hi Lucian, I just had a look at the spec. Yes, namespaces must stay, but many links can be updated. Here my impression:
Replace http://www.biopax.org/ -> https://www.biopax.org/ (url, https exists) http://www.uniprot.org/ -> https://www.uniprot.org/ http://identifiers.org/ -> https://identifiers.org/ http://www.ebi.ac.uk/ -> https://www.ebi.ac.uk/ http://www.genome.jp/dbget-bin/www_bget?ec:3.5.4.4 -> https://www.genome.jp/dbget-bin/www_bget?ec:3.5.4.4 http://www.cellml.org/specification -> https://www.cellml.org/specification http://tools.ietf.org/html/rfc6350 -> https://tools.ietf.org/html/rfc6350
Remove http://www.godatabase.org/cgi-bin/amigo/go.cgi?view=details&query=GO:0007268 (link is dead, resource probably does not exist any more) http://us.expasy.org/cgi-bin/nicezyme.pl?3.5.4.4 (link is dead, resource does not exist any more) http://www.math.uwaterloo.ca/bingalls/Pubs/conservation.pdf (link is dead, resource does not exist anymore) . Not sure about http://biomodels.net/sbo/ -> https://www.ebi.ac.uk/sbo/main/ (forwards, https exists, careful this is also part of namespaces, only replace when the website is mentioned)
Don't replace http://sbml.org/ (urls, no https yet) http://www.sbml.org/sbml/ (namespace) http://www.w3.org/ (namespace) http://www.mysim.org/ns (namespace) http://www.struct.org/ns (namespace) http://www.othersim.com/ (namespace) http://www.sbgn.org/ (url, no https yet) http://www.psidev.info/index.php?q=node/60 (url, no https yet) http://biomodels.net/ (used in definitionURL) http://co.mbine.org/ (no https yet, part of namespaces) http://purl.org/dc/ (namespace) http://www.informatics.jax.org/searches/GO.cgi?id=GO:0007268 (url, no https yet) http://dublincore.org/* (url, no https yet)
Hope this helps. M
.
Original comment by: matthiaskoenig
Original comment by: luciansmith
Accepted and pending for l2v5. Also added to http://sbml.org/Documents/Specifications/SBML_Level_2/Version_5/Collected_Errata_to_the_SBML_Level_2_Version_5_Specification
Original comment by: luciansmith
As these changes have been made to the source code, following our new procedures, I'm going ahead and closing the issue here, as we no longer have a 'pending' status between the issue being addressed and the spec being released.
Apply to SBML Level 2 Version 5 release 1 and SBML Level 3 Version 2 draft (August 2014)
In page 105 of the L2V5 specs, there is an error in one of the example of nested annotations. Identifiers.org url should be used and more importantly the namespace used for the protein modification ontology is wrong (it is not 'obo.mod' but 'psimod').
Here is the xml in the specs:
<species id="S1" metaid="_000004" compartment="cell" hasOnlySubstanceUnits="false" boundaryCondition="false" constant="false">
It should be changed to:
<species id="S1" metaid="_000004" compartment="cell" hasOnlySubstanceUnits="false" boundaryCondition="false" constant="false">
The same problem happen in the L3V2 draft, on page 108-109
Thanks
Reported by: niko-rodrigue
Original Ticket: sbml/sbml-specifications//293