I'll try to go quick. So thanks for this opportunity to give this talk. So I'm going to talk a little

bit about CERN and the HEP software foundation and the software life cycle that we're trying

to get people to follow. And so just as one slide introduction, probably most of you have

heard of CERN. It's home of some of the biggest scientific experiments. The main one is of

course the Large Hadron Collider with its experiments, ATLAS, CMS, LHCB, and ALIS. But

then there are also a lot of smaller projects that also happen at CERN. And since CERN has

a Large Hadron Collider, it's also kind of the central hub currently of high energy physics

worldwide. And so it kind of has place on rolling coordinating also community wide activities.

And so you might have seen these huge hardware detectors that are the biggest kind of experiments

that we've ever built. But obviously besides this massive hardware, everything we do is

informed by software. And as in scientific research, generally obviously we spend a super

wide range of different software types. So some type of software that we produce is almost

product like. Like ROOT as a data analysis framework or a Gen4 as a simulation framework

that is also used in commercial applications and whatnot. And then kind of go through the

entire gamut of types of software where you have different kind of levels of professionality

and it kind of ends with one-off macros to do some specific plot or submit some job into

the batch system. And so all of this kind of needs to be managed. And in our community,

we have a pretty big challenge ahead of us, which is called the High Luminosity LHC, which

is kind of the last 20 years of the LHC. And there we kind of estimate that in the current

funding structure, we will kind of have a hard time with our software if we don't do

a lot of R&D in our software development. And so all the software across this entire

spectrum kind of needs to rise to the occasion. So we try to equip the community with the

right infrastructure and the training to kind of rise to this challenge. And so the HEP

Software Foundation was founded in 2015 to kind of help tackle these challenges or to

provide a forum that is not tied to any one institute or one grant or something and to

share ideas and experience and also to encourage best practices. And so the discussion there

is happening in various working groups. So we have working groups on data analysis, detector

simulation, frameworks, and data acquisition software and all that stuff. And so one of

the main outcomes of the HEP Software Foundation was the community white paper process where

we have a pretty detailed roadmap of what we need to do in the next 10 to 20 years in

terms of software and computing. And so obviously this is a lot of material which is hard to

summarize, but then I want to highlight two trends. So there's like one trend to kind

of integrate with standard industry tools and hydro physics. We've always been ahead

like a first mover's disadvantage to a lot of the degree where we hit certain problems

computing very early on before the wider world hit them. And so we developed our own tooling,

but now that big data is everywhere, we have the opportunity to connect to the standard

industry tools and this kind of lowers the burden of the stuff that we need to maintain.

And then we also want to make our own software more modular and interoperable between different

experiments so that we can kind of mix and match the software that we're writing and

not everybody is writing their own huge monolith. And okay, so we've seen a lot of these kinds

of recommendations, so they're basically the same for us. So we also kind of suggest that

people kind of have a license, right? And so the licensing has been traditionally a

little bit undervalued, but if you have your code on GitHub, it's extremely easy to add

a license. So you should really try to do that because it helps make your code reusable

and people can safely integrate that into their project. And so if you do this or try

to do this after the fact, it's very difficult because then you need to kind of get permissions

from everybody. So it's a really good idea to do this from the start. And yeah, there

are various legal reasons, but I heard from Katharina that there's going to be a dedicated

talk on this. So then contribution guidelines. So again, GitHub makes it extremely easy to

add these things in a kind of standardized way. So a lot of times software, you know,