May 2023 Archives

PTS 2023

By Tux on May 9, 2023 3:45 PM

Asked to blog, my first reply would be "I do not blog". And I don't. Usually.

Reading the reports from the other participants of the PTS 2023 in Lyon, I however see the value to those that did not have the chance to participate.

Some history

Being the only participant that attended all previous events too, some history might be to the point.

This event started in 2008 when is was organized in Oslo with no agenda at all. The organizers (Salve) had the vague idea to get people from the perl world together in a meeting and discuss "things" that were related to "Quality" and "Testing". The organizers invited people that they thought to be of value and involved enough to get somewhere. Here are a few lines of that initial mail:

Subject: Would you come to a QA workshop/hackathon in Norway in February? Date: Mon, 19 Nov 2007 13:25:04 +0100

I had a chat with Adam Kennedy on IRC a few days ago. There I proposed to arrange a QA workshop/hackathon in Oslo, Norway. Would you be interested in coming?

The purpose of a QA workshop would be to solve some problem that's easier to solve when everyone is gathered in the same physical location. :)

Nothing has been decided yet, and there's no funding yet. This is just an idea. Don't get your hopes up too much.

The workshop must be in Oslo, and in relation to another conference about the use of free software in the public sector/government. The details of the conference aren't set yet, but a tentative date is sometime in the last week of February 2008. The reason for this is related to funding.

Everyone who comes has to give one talk at the technical (or non-technical) track at the conference. Pick a topic that interests you, but it has to be relevant for the conference somehow.

Who should be invited? You tell me. …

What's the workshop topic? You decide. …

Is this a CPAN/Perl-only event? Preferably not. …

Al least to me it was just that. When we arrived, that was the introduction, and the agenda was a clean flip-over board and we were asked to put on some topics to discuss. The event was baptized "Quality Assurance Hackathon".

In the first two events (Oslo 2008 and Birmingham 2009) we discussed TAP (Test Anything Protocol), parallel testing in CORE and the before-base of CORE: Configuration, what is CORE, what is Test, where do tests start and how can we get reproducible test results. This is also the period where Test::Smoke was embraced by CORE.

Then 2010 Vienna, 2011 Amsterdam, 2012 Paris, 2013 Lancaster, 2014 Lyon, 2015 Berlin, 2016 Rugby, 2017 Lyon, 2018 Oslo and 2019 Marlow.

Lancaster was probably the one with the greatest spin-off, as a lot of decisions were made resulting in The Lancaster Consensus and most of that event was discussions.

Then came Covid and all went silent.

The organizers asked the core members of the last few events if they wanted to organize yet another when it was possible again to meet and most of us - including myself - replied with "yes please", but none of us were able to dedicate the time and effort to organize it, so it went silent again.

Until BooK stood up and just did it. 2023 in Lyon.

Participants

To understand how the group that attends is gathered needs some explaining. Unlike most gatherings like conferences, workshops and open hackathons, the main focus of the Perl Toolchain Summit is the perl toolchain: from Configure (which builds perl itself) to cpan (used to install new stuff from CPAN) and all in between. It may not be a surprise that every part in there should know about the other parts so the end-user experience is as smooth as possible, so the process is like this: the organizers select 10 people: those that have the most impact on the different parts in this chain. Those 10 are asked if they want to participate and if yes, in which period they are willing to come. A few weeks later, the date is set, and the people of that inner ring are asked who they want as next 10 to attend. They choose people that are both knowledgeable and motivated and are likely to have a positive effect on the results of the meeting given to open issues at hand. The suggested people are then put in a spreadsheet where all 10 can vote and the voted-for people in this second circle are invited. It may be clear that none of these first 10 attendees are set in stone, but with some people working is this area for the past 25+ years some names are obvious. The now formed group of 20 may suggest yet another round of people to invite and the 20 vote. The final group is set to ±30 people as that has been proven to be the most effective.

This process has resulted in over 100 different people having attended this event over time!

I've been to all sorts of meetings, but the PTS (formerly known as QAH) is the one that needs the most and most thorough preparation. We are expected to share our agenda so others can prepare to chime in. There are many overlapping areas and getting together means we can also sit together and stare at the same screen and call a third or fourth pair of eyes to look for solutions or spot problems.

On the first day we all stand up and tell what we plan to do. People raise their hands if they want to take part in a discussion. Then people find a place and start working. Some alone, some in pairs and some in larger formations. Discussions happen all the time and that interrupts "programming".

Agenda

My personal agenda was two-fold:

Make Configure 100% in sync with metaconfig again. Some 4 years of scattered patches, changes and non-coordinated rumble had to be sorted
Get motivated again. Covid seriously impacted my motivation to work on perl5 core issues. Motivation surely helps with point 1.

Results

Having all CORE people present, was very useful to make decisions I usually have to make on my own. Configure is not the CORE area many people enjoy as it involves a second repository and aditional knowledge of which very little has to do with perl. So point 1 was just doing a lot of work and having no $work distractions helps a lot.

Next to that I helped Abe with Test::Smoke and revived some smokers that were either misconfigured or configured in a way that they did not yield useful information.

I seriously enjoyed a bouldering event that was organized as time-out. It was 40 years ago I did that and I was pleased with how that went.

Looking back after day 2, point 2 could be marked as PASS :)

The last two days I had a go at a new module (Test::CVE) that has not yet been released. It is based on discussions about EU regulations (NIS2 & CRA), CVE reports and other known vulnerabilities.