XSF Discussion - 2020-09-06

If you are an author and you care about your implementation being listed, making a yearly reminder is not a very large hurdle.

Ge0rG: "shit happens" and yearly reminders are easy on failure

Maranda: make a cronjob that will `echo "renew xmpp implementation list"` and there is your yearly reminder email

Maranda: also we don't even know who the authors are

That's a bullshit and you know it, the most, you don't have a mail contact field in the data

Maranda: so do you have a better idea of proxy measuring whether the author is still interested in xmpp, that has the minimum of additional work burden on the XSF web team?

Ge0rG no I measure, that courtesy is a good thing, and that the author like myself may simply forget he has to renew the entry every year while caring about the project. So sending a mail out when an entry is phased out is not a bad thing and makes the whole organization looks less like it is run by a bunch of assholes to be frank.

That's my thought

Maranda: this was also discussed before, and IMVHO, an author that hardly cares about the XSF would be likely to use such a reminder email to renew their listing, even if otherwise not doing anything to keep up with the progress of xmpp.

So requiring the author to care a little bit on their own is not so much asked after all.

Also this listing is not for your ego, it's for the users looking for viable implementations. That might better explain the process.

But it seems to not work as intended as we recently saw that a lot of active projects were not listed anymore while projects like pidgin were.

Ge0rG: again that's BS, who said anything about ego and what makes you presume I don't care? I even probably had a reminder somewhere, but, this year I just basically moved the whole infrastructure to new VMs/Hardware and Software and I won't mention the other burden starting with a "C". So you'll excuse me if _the reminder_ got lost in the various transitions.

All to say *shit may happen* as mdosch implies, and this is the kind of absurd, wrong elitarian attitudes that causes people to scorn on the XSF. ✏

Maranda, it’s not about you, it’s about pidgin et al. ;)

but what mdosch says is indeed relevant

Maranda: so your argument is, a reminder that you set yourself will get lost, whereas an email reminder from some random machine on the internet won't?

No not at all, I think I spoke clearly. I said that if you phase out a software from the page you could at least pay the courtesy to inform the author. There could be many reasons as of why the entry wasn't renewed, mine did expire in July and while I follow the XSF activity, I found out just now

It's maybe just me but I find that phasing out without even a word, is quite rude. And since you do send "reminders" in the case of quarterly membership renewals I don't see why you couldn't act similarly

FWIW I don't get the downside of email reminders either. Specifically when it's not about the software author's ego but about the XSF's interest in providing a sane list (i.e., we want something from software authors, not vice versa).

> a reminder that you set yourself will get lost, whereas an email reminder from some random machine on the internet won't? Maintainers might forget to set a reminder, or it might break for some other reason. The idea seems to be that we only want to list software where maintainers care enough to be listed. Whereas I would've thought the goal is to list software that's maintained.

also we are always talking about maintainers, but everyone could do the update or not?

for example im maintainer for some years now, but i dont remember doing an update ever

so someone did that for me it seems

Yes that's how I understood it.

So basically it goes on popularity? Even?

I did that for Gajim. but commit history reveals Ge0rG did it for lovetox as well :)

I don't have an issue with anyone submitting a renewal, but this was specifically rejected a couple of months ago. We require a verifiable maintainer of the project to request the renewal.

Ah.

Will need to dig into chat archives / meeting minutes to find the specifics.

nice, it would totally work if Ge0rG does it for me though, i dont need a reminder then

🤣️

I don't expect we will ever find a mechanism that will 100% satisfy everyone. I do believe that what we have now is an improvement over what we had before.

We can obviously tweak and adjust as we learn the effects of what we now have in place.

As the reminder it could even go in the form of a message on (one of) the (mostly dead now) mailing lists thrice a year, I don't see that as much the "additional work burden" for the web team Ge0rG talked about imho

However, what we have now is the result of quite some work and a lot of discussion. Dismissing it offhand as "bullshit" is to easy.

Dismissing a possibly valid argument by reducing it to the b* word kinda sounds easy as well 😛️

And as far as I'm concerned I can always relate to something not being done due to missing manpower. I just didn't quite get the "it's the maintainer's responsibility to be listed; not our problem" stance.

and in the past discussion it was decided that the goal was to actively *not* remember people about the update?

like its intended to be a little challenge for maintainers?

sounds weird, when i remember this correctly it was just so dead projects dont get listed anymore

so a maintainer coming in and doing the work is ok and i think nobody challenges that

Guus: the BS were the arguments used to counter my own and... actually *dismiss 'em* tbh. Beside I'm the one who pointed the issue..?

its the "But we dont remember you about it" thats weird

I've not made my mind up either way. I do wonder if sending out reminders defeats the purpose of this mechanism, which is to purge the list of crappy entries.

Guus, i dont think thats a stance the XSF can take

whats crappy and whats not

as the software site explicitly says, that the software is not tested on compliance or anything else

Hence the renewal mechanism lovetox

so XSF intention is not to list "good" software whatever that means

the intention is to list "active developed" software

It was assumed that by requiring maintainers to renew, we'd weed out the old, unmaintained, crappy listings

yeah and i think thats still not wrong

Guus again it's just 3 mails a year, nothing differs. Beside a better attitude torward the developers by the XSF

Remembering them to renew might defeat that purpose, unsure.

That's all

but i dont see how a once a year email to remember people to update would counter that?

I have no strong feelings either way.

For me the incentives of being on that list aren't big enough to go through the troubles of renewing

The xsf probably benefits more from listing Conversations than I benefit from the traffic

That's why I don't like the requirement of needing a maintainer of the project to do the listing

(judging from the number of clicks that github tracks as coming from that site)

I think that you have a valid point Daniel .

I do hope that in your role as an XSF member there's more of an incentive for you to have Conversations listed, Daniel 😁

Yes. Luckily someone did it for me this time.

I need to go, family commitments. If we want reminders, let's involve web team/iteam.

Daniel: there are people who think that it's good to have pidgin listed there, and I'm very strongly convinced that listing it is harmful to the xsf

Popularity is a horrible proxy.

Indeed, beside if you let anyone do the update it may go against the will of the mantainer in the case he/she doesn't want his/her work to be on that list ✏

Everything will be much better once we have compliance badges

Ge0rG: why not rank clients by how conformant to the compliance suites they are?

Let's not put an insane amount of effort into this. The list will never be 100% up to everyone's liking anyway. What we have now is a massive improvement over what we had. Let's just call that a win.

Maybe it's time to put on the table getting rid of these lists ✏

There's something to be said for that. It feels kind of awkward that the XSF apparently feels tie the need to publish lists like this. As a protocol with a very mature ecosystem, that perhaps isn't needed?

I really am off now though.

eta: that's a great idea.,but how do we get the required data?

Ge0rG, I guess we’d need infrastructure anyway if we plan to make use of DOAP processing (automatic client badges for the list, listing example implementations in xeps, rendering supported xep tables for clients, ..)

wurstsalat: yes, that's important but also complex work that's looking for a volunteer.

right

There will be questionable claims of features from projects if having those features listed gets them a better position.

Guus: well, we can't prevent people from lying, and it will be a very interesting exercise in xsf neutrality once that happens and is contested. But given the developer interest for the xsf lists, I don't consider that attack vector as very likely

Guus, sure, but I guess there is no technological solution to that

I've already thought about how to sort such a list. CS year first, advanced / core second, category third?

There really is no "right" sorting order for a three dimensional data set

yaxim is 2020 Core IM, Advanced Mobile. What would dictate the sorting priority?

Maybe just listing the supported badges is just enough?

(this is also why I used the colors to indicate if the year is current or not on my badge prototype)

Ge0rG, make the table sortable (like the xeps table)

wurstsalat: sortable by what?

Year and Core/Advanced + filter for mobile/IM ?

So by all three dimensions? Separately?

I think 'Year' would be the important one. if you’re looking for mobile clients, you probably want to filter by Mobile category and then have a list of 2020;Mobile (Basic/Advanced)

So we need filters in addition to sorting?

Maybe we are overdoing it now

Or maybe we first need to fix the tooling to actually support fetching DOAP

Ge0rG, sure, without the tooling we’re just bike-shedding :) but the xep list has filters as well, so we might have a look there

wurstsalat: you just volunteered?

I volunteered several times, but working on the website is a pain at the moment

I'd rather keep it all a lot simpler. What we have now, with some tweaks maybe, or complete removal of the listing. Most other options won't improve things dramatically, while take a lot of effort to create and maintain.

Guus: I see the value in listing supporting (open source) implementations in the XEPs from DOAP data, and the additional win of compliance data in the listing would actually improve the overall situation.

Also I'd be okay with listing pidgin, if it comes with a "Core IM Client 2006" badge

Compliance Suites are dating back to 2006 ?

https://xmpp.org/extensions/xep-0211.html is 2008