Still needs review:

• external_identifier
• have the author in a CodeMeta field with affiliation?
• releaseNotes
• url (for origin)
• add downloadUrl? (https://www.ipol.im/pub/art/2018/236/mlheIPOL.tgz)
• is there a notion of version

@ardumont can we have dots in a deposit slug / external_identifier?
An example:
ipol.2018.236

Here is the new IPOL proposition (made by José-Luis):

<?xml version="1.0"?>
        <entry xmlns="http://www.w3.org/2005/Atom"
               xmlns:codemeta="https://doi.org/10.5063/SCHEMA/CODEMETA-2.0">
        <title>mlheIPOL</title>
        <external_identifier>2018-236</external_identifier>
        <author>Jose-Luis Lisani</author>

@vlorentz what do you mean with free?

Here is the command:

thanks ! for the feedback..

In T2381#43854, @vlorentz wrote:

swh-loader-tar (used in the revision creation with a field called extrinsic with two properties provide and tool)

I don't think we should change this right now, but convert them when we have a proper storage for extrinsic revision metadata

Changing provider in title to authority for terminology consistency.

I suggest the following changes:

Could you change the trst with two different persons? for a better coverage of changes?

I have updated my first comment with maybe a more clear and without typos sentence.
Sorry about that.

@ardumont After this diff lands as is, the pattern in loader with the extrinsic field, needs to change terminology to authority and fetcher.
(instead of provider and tool, to stay coherent)

just worrying over branches ;-)

Can you show me the result of an end to end test locally with a real XML to see the result in the revision metadata and in the origin_metadata ?

Thanks for the fast diff !

:-D
I'm on it..

Good !
seems we are converging on this.

If I understand correctly this is the concatenation of two different models:

• one created for the deposit
• one created for PyPi loader

The task description says that the property extrinsic should be deleted in the revision metadata.

Here is the first changed revision: https://archive.softwareheritage.org/browse/revision/c33910a29d53f4e137c225b21a8d59e43327cbf9/

I don't mind calling it change :-)

One difference between HAL and IPOL is the item identified with the identifier:

• In HAL, the HAL-ID is given to the software artifact alone and if you want to deposit the main software article it will have a another identifier.
• In IPOL, the DOI identifies the metadata record of both the published article and the published code

Also the IPOL article and code are peer reviewed which results with a publication with an IPOL DOI.

Could you do a test for a registry and a forge, for the same origin?
I can give you real life examples, if needed..

From the article web page:

published
    2018-12-07 
reference
    Jose-Luis Lisani, An Analysis and Implementation of the Shape Preserving Local Histogram Modification Algorithm, Image Processing On Line, 8 (2018), pp. 408–434. https://doi.org/10.5201/ipol.2018.236

I will continue my review later tonight.

The extrinsic property seems like a bug. needs further analysis.

Just a question about using a path with a different branch, for example for a tag of a version (which is not a release):

• in this case, the anchor is the snp and the branch name (the tag) is in the path?

Here is the link to the open issue on the CodeMeta repository:
https://github.com/codemeta/codemeta/issues/240

A few suggestion for the repository view:

1. I like the divided tabs, but in each tab, the context should be kept
2. In source code tab : context of the source code (SSC)
   • opencv\opencv\[dir-id] with snapshot date and branch
   • Also readme should be kept in the SSC tab
3. About metadata:
   • there is metadata in revision
   • there is deposited metadata for origin in the origin_metadata table
   • there is indexed metadata for origin
   • there will be metadata that can be associated to each artifact
   • A metadata tab is nice but will be very difficult to understand to which part of the repository the metadata is relevant,
     • show metadata as a separate button, as proposed in the task, is a temporary solution
     • we should reflect on this with the upcoming metadata additions (most registries, show metadata on the right, with several metadata functionalities, see Zenodo, HAL or SwMath)

seems so from the design ;-)

try again button, will it appear only for failed requests ?

The behavior of the page:
I enter the link to the repository
click submit (which I suggest to change to save)
and there is a green box appearing below the form.

I think the reference here should be the vault page

From the design, there is also a change of colors on the visits page.

Great UX proposal not to interrupt the browsing with the download page.

Will you change Visit type into Origin type?
the former is more accurate, the latter is more comprehensible by someone that is not familiar with the data model.

I agree that a search box is a great way to keep the user engaged.
The text in the search box can be leading, here in the example we have Enter keyword, but the result is a search on the url, right?
If the search is the actual search on url:
Enter url
Enter repository
Search the archive

In the sidebar menu I would keep the API access (unless it is accessible during the navigation - not on the landing page).
Also the access to the API on the side, is very small.

I've seen the new example, is this the right transformation?

If you agree with this proposal with @vlorentz's example, I can submit the issue tomorrow.

Great news !!

I just discovered that there is a pending term maintaineron schema.org, thanks @vlorentz
for pointing this out.
https://schema.org/maintainer

@vlorentz can we say that your new validator resolves this?

Here is the first draft:
https://www.softwareheritage.org/?p=21219&preview=true

Available here:
https://codemeta.github.io/codemeta-generator/

Thanks @olasd for the accurate and detailed summary !

Thanks @anlambert ! That's great news !!

After discussion with @rdicosmo, GitHub is the best place for the community that wants to use the CodeMeta Generator.

I actually don't have it in both cases.
I'm firefox quantum 68.2

Complete email: