Nowadays, installing QGIS seems like straightforward : download the installer, run it, click "next, next, next", and enjoy ! While this is true for common single users, the situation can be very different in large organizations. With QGIS becoming the de-facto desktop GIS, and adding server capabilities, large deployment have become a strong need and a reality.

In big corporations or large public organizations, the IT infrastructure is usually complex, and requires a specific attention to the way software deployments are achieved. Installing QGIS for hundreds or thousands of users can be a major challenge.

The following technical issues, among others, have to be tackled :
- Automated installation
- Dealing with multiple user profiles
- Automated configuration
- Dependency management
- Restricted Network access
- Security
- Customization and packaging
- Upgrade management

Moreover, outside of technical aspects, organizational aspects have to be taken into account. Support, helpdesk, maintenance, upgrade policy, training, funding…

We will show real-world examples of QGIS deployment in big corporations and large public organization. Based on our experiences with QGIS large scale deployment, we explain in this presentation how we dealt with all these topics, and the lessons we learned on the way.

Terristory is a web platform providing a sustainable energy observatory oriented towards decision-makers and territory planners.

AURA-Energie Environnement is a French association acting on account of the Auvergne-Rhône-Alpes region to promote sustainable energy. Aura-EE started the Terristory project a couple of years ago, in order to put their data on the web.
Energy data is geographic by nature, and one of the main aspect of managing energy is being able to observe its characteristics on a given territory.

From a simple data viewer, Terristory evolved into a full platform for data observation. Dynamic graphs have been added, and some advanced features like :
- create scenarii on Energy equipment ( e.g. build a methanizer )
- impact of decisions on local employment

Terristory is based on OpenSource software : PostGIS, Python, OpenLayers, Vector tiles… The full code for the Terristory platform itself is opensource and will be published publicly in 2021.

Terristory was initially funded by a single actor and deployed in a single region. In 2020, the project accelerated : it evolved into a consortium to support the platform and deploy it in other regions. This evolution made Terristory a national project, and a reference platform for energy data visualization. This mutation is interesting on multiple levels, as it is totally coherent with an opensource project :
- from a simple project to a full platform
- from a single developer from a single company to multiple developers from various origins
- from a single funder to multiple funders organized as a consortium
- from a single actor for roadmap definition to a mutualized roadmap

This transformation makes the project's history and experience unique. The battle for climate is open, and platforms such as Terristory have a strong role to play. It should be an inspiration for any project oriented towards opensource, opendata and resource mutualization.

Sometimes data is not enough, and we need metadata : we want to know how and by whom it has been produced, altered, updated.

We also want to track history of the data, be able to go back in time, and even sometimes deal with different versions of the same data.

There are numerous technical solutions to cover these needs. We will present some opensource solutions and the associated ecosystem : PostgreSQL mechanisms (triggers), qgis-versioning, pg-version, fastversion and others.

We will speak about different possible use-cases for data versioning and which technical solution is the most adapted, from simple data timestamp to full-fledged history needs.

The talk will be illustrated with some real-world cases.

There are some new GIS kids on the block : pg_tileserv and pg_featureserv are GO components allowing you to serve GIS data to build cartographic web applications.

Pg_tileserv let you take any data stored in PostGIS, and easily stream vector tiles to the web.

Pg_featureserv also takes your data from PostGIS. It implements the latest OGC API for Features ( OGC APIF ) standards and make it simple to provide a interoperable API for your front-end application.

These components take full advantage of the underlying PostgreSQL/PostGIS database to provide efficient features with a very simple mechanism.

Easy to set-up, easy to use and powerful, they are now part of the de-facto stack for GIS web developers.

We present these components and showcase an example to demonstrate their simplicity and efficiency.

Innersource is the use of open source practices... within companies for their own and internal needs.

Software development is no longer an ancillary aspect of the activities of large companies and many actors believe that "classic" project methodologies lead to failure and are a source of suffering for the teams.

As a consequence, the innersource movement is rising, and is supported by the spread of so-called agile methodologies because it is a broader reflection on "how to produce sustainable, quality IT projects".

What are the objective and quantified reasons that lead companies to adopt innersource approach ? What is the expected return on investment ? What are the limits of the approach and its links with open source in general ? How is Innersource positive for the OpenSource movement ?

We will present how specific companies put innersource into action, as a few of them allow public communication of their program. The approach is wide-ranging because it deals with technical aspects, but also, and above all :

• governance of projects
• human organization necessary for its operations
• legal aspects of ownership and IP
• communication carried out by the company

Innersource programs also provides feedbacks on successes and failures, improvements needed and key issues. These experiences can be very useful for opensource communities to improve their own processes. We will broaden the presentation by showing how the choice of IT project methods translates into human systems.