A new type of Software Directory?

I think we need a more durable, scalable, democratic and decentralized type of directory.
I fond a project with a .yaml file describing software. Such a structure seems like a better way to run a software directory than what we have now. publiccode.yml
The FSF Directory uses a semantic wiki which I totally don't understand.
The future I imagine is

  • standardized structure for describing software
  • standardized API with decentralized data
  • public access to add, search and edit entries
  • easy wrapping in websites and other presentations

I don't really know how to move this forward, it seems to me that the two projects above must be worth more investigation. Sadly I haven't been able to make time to look into them further.

AceHobbes

Comments

  • edited June 2022

    I like our current wiki directory. Positives things are:
    It is a limited selection of apps with potential for AEC industry.
    It shows in a simultaneous way (in a single page) many tools and it is friendly with mobile phone browsers.
    It highlights tools with limited exposition between AEC users.
    It classifies apps in AEC workflow stages.
    It associates several apps in each groups, so people can explore interoperability and build possible workflows.

    I consider that any improvement in the Directory should try to keep these features, because they are very useful for final users.
    Maybe this directory could be improved to made it more useful for developers and potential collaborators?

  • edited June 2022

    All of the things you mention should be achievable with a page built on well structured data. What we have now is very manual and unstructured. We should be able to call a database for things " 'Usergroup' includes 'AEC', 'License' is 'gpl2.0-only', 'category' includes 'Finite Element Modeling', name, icon, screenshot, URL, short description, long description, resource list, operating system, version, maturity, ... " and just present it the way we want.

    What we have is not even very searchable, not very organized by category or function, and I think if you look around at other websites with lists you'll agree that lists can look much better and function much better.

    Examples
    https://alternativeto.net/category/phots-and-graphics/cad/
    https://aec.works/companies?sort=updated
    https://www.edrawsoft.com/list-of-best-drafting-software.html

    How about changing your avatar image - you always look so sad :-)

  • edited July 2022

    I've made a small improvement to the Software InfoBox which means that the maturity level of the software now adds that software to the appropriate category.

    For some reason this change to the InfoBox template first adds it to the category next time the page is saved. So the software maturity categories will only slowly be populated. To fix this in one go just go to the software category, open all pages in new tabs, edit->save.

    https://wiki.osarch.org/index.php?title=Template:Infobox_software
    https://wiki.osarch.org/index.php?title=Software_maturity
    https://wiki.osarch.org/index.php?title=Category:Software

    Hope this is helpful @Martin156131

    Martin156131
  • Found this interesting solution ... https://www.wikidata.org/wiki/Q8038

  • Not an easy task...

    From my point of view, the standardized structure for describing software is the key point. Some times we want to have as much information as possible... and forget that we will need to gather and process a lot of information. That could lead to a lot of voids in the database fields (lack of time or comprehensive information). I mean, it's needed to pay special attention to the database fields (neither too much nor too little info).
    I don't know if the basic structure is already set/decided. If is not the case, it can be a good starting point.

    Other free software directory (as example); https://framalibre.org/

  • @Hobbes said:
    Not an easy task...

    From my point of view, the standardized structure for describing software is the key point.

    The idea is to have a suggestion for how it can be done and then talk to other projects like

    Then getting as many projects as possible to agree on a good solution. Then all those projects can pull from the same data to populate their own lists. If they only want to list mature projects from the 'vector graphics' category that run on windows & linux. They can do that.
    I'm looking for a way to crowdsource a FLOSS software list instead off all this poorly structured data and duplication of effort.

  • So, if I have understood correctly, the idea is to have a common software database as source for all those (or other) projects.

    An excellent idea and very ambitious project (even more complex than I was thinking). But you're right, there are a lot of duplicated efforts and poorly results.

    @duncan said:
    The idea is to have a suggestion for how it can be done and then talk to other projects like

    From the point of view of coding?

  • As far as I can see we could just start by talking to the team at https://www.wikidata.org/wiki/Q8038 and adopting their structure if we can. Then we can just start improving their project. We may need to have a discussion about the bar for notability which they have (fair enough). Do you also think the structure is usable?
    Of course the obvious alternative is to adopt the semantic wiki system of the fsf directory. But they're not easy to work with and the system looks pretty messy.

  • I don't know the structure of wikidata in detail, but I think is usable. At the end, the information of a directory is quite simple compared with the scope of wikidata (any type of data), so it's not bad idea starting from their structure and trying to simplify it.
    Not sure about the concept of notability. If it's a way to measure the reliability of the sources, from my point of view it's not a critical aspect (the relevant info will come from the software developers).

    Regarding semantic wiki... I'm not sure how it works (I think it's more suitable for other type of data). Anyway, I agree with you, looks messy.

  • @Hobbes I think some data scraping from wikidata to our own site is the way to do it. Some basic scripting can fetch the data we want to present.
    There might be two places using the data

    1. Software Directory: an automated list fetching minimal data for everything meeting relevant criteria. Using that we can manually link whatever we choose to our own pages...
    2. Software Pages: That could hold wikidata in some kind of box (or two) on each software page. Then we can manually write a more subjective text contextualizing the software.

    By doing this we contribute to maintaining a centralized list anyone can 'fetch'. We also gain from the work of others as each software wikidata page develops.

Sign In or Register to comment.