PROJET AUTOBLOG


Planet-Libre

source: Planet-Libre

⇐ retour index

Simon Vieille : Réseaux sociaux : flux à lier - #DATAGUEULE 95

mercredi 26 février 2020 à 08:31

Ils sont attirants, attachants … et rapidement indispensables. Les réseaux sociaux nous appâtent à coup de likes et nous bercent dans leurs "infinite scroll". Et rapidement leurs interfaces nous poussent à la consommation jusqu’à l’overdose. Et ce n’est pas un hasard. Ils ont bâti leurs empires sur notre addiction à la dopamine.

Retrouvez toutes les sources de l'épisode sur : https://wiki.datagueule.tv/Réseaux_:flux_à_lier(EP.95)

Gravatar de Simon Vieille
Original post of Simon Vieille.Votez pour ce billet sur Planet Libre.

Articles similaires

Francois Aichelbaum : Agile and Infrastructure

lundi 24 février 2020 à 08:11

Recently, I had a rather interesting discussion with some of my clients over Slack: Agile processes applied to Infrastructure. At the beginning, we did not reach a common ground, but we quickly realized that the problem was not our approaches, but our vocabulary. Agile? Infrastructure? Once we agreed on these terms, the discussion took on a completely different form. Here is a short summary.

Agile ?

Often, when we talk about Agile, we talk about agility. It is assumed that it is about being flexible in our actions and reactions. It is mainly a mix between a mindset, a culture, an approach and above all a paradigm.

Defined in the Agile Manifesto, this approach to project management is the opposite of traditional approaches such as the V Cycle or cascades. The short version is that these traditional methods lead irremediably to a tunnel effect, which can have disastrous effects on both the project and the client relationship. Conversely, the Agile approach proposes to reduce this tunnel effect as much as possible in order to allow better transparency, through an iterative and incremental process.

So the key elements are there: transparency and iteration. From this, many other concepts will derive, related to the different methods but also to the contexts of applications. Many sites and books will describe them much better than a few lines on this blog.

Infrastructure ?

An infrastructure is a set of interconnected elements that provide a framework to support the entire structure. As such, it is not just the network or system, but a collection of these elements and many more, to be considered in context. In the case of a team, we will have to talk about, for example, NetOps, DevOps, SRE, Architects, Infrastructure Engineers, … All working on the same subject, with different angles, and therefore different needs and means: their daily tasks, depending on their position and the context, will range from pure Run, to R&D applied to infrastructure.

And so ?

Okay, so let’s go back to our original discussion. Why talk about Agile and Infrastructure? We had different points of view together (I arrived during the discussion) and the central point was about the ability of “Infrastructure” teams to work in an Agile way.

One of the participants, who manages an operational team, therefore pure Run, did not see how Agile could be applied to the Run. The other, an R&D manager, was suffering from blockages from an Infrastructure Engineer who had joined his team after years of working on Run. The last one was managing a team that did both R&D and Run, and especially managing a rather junior DevOps that he was happy to model as he went along.

Depending on the interlocutor, it was more or less possible to apply the Agile approach to Infrastructure and that if someone couldn’t do it, it was simply their fault. Rather abrupt.

In fact, we must remember that in the professional world, it is our experience that models us. So you have to look at the current tasks to be performed but also at the past tasks of the people who will be working there. A person who has never worked in project mode will potentially find it more difficult to integrate into a project mode than someone who has been in it since the beginning of his or her career. And in Infrastructure, there are a lot of opportunities to work in project mode, but also a lot not to.

Making the infrastructure agile

So the real point is there: apply the Agile approach only when necessary, on the project, with people who have an approach to the subjects compatible with the two key elements.

Indeed, we could try to apply Agile methods to the Run, but this will not concern the core business rather than more related topics (documentation, debug tests, …). In the same way, a person whose experience has taught him how to look directly for the right final solution, even though he will spend a long and unmeasured amount of time on it, will not easily be able to work in incremental iterations, more time-limited.

This is the point that I sometimes have fun stygmatizing with my clients to trigger reflection. I then caricature by talking about the American approach of a startup, ready to do anything and question itself the next day with a complete turnaround; versus the French approach, much more posed, and which will only deliver once perfection is reached.

So it comes down to the fact that we can work in project mode, according to the Agile approach, with teams that are used to working iteratively and incrementally. The basics. In reality, therefore, we can come up against other problems:

On the first two points, an Agile coach is the best way to find an answer. Indeed, the Agile coach will make sure that both the mindset of the team, but also the coaching structure allows this work in Agile. In the same way, he will often overflow on the means made available. He will then continue the work on the method, because in the Agile approach, each implementation comes with its own method and its own framework that will serve as a basis for its integration. The Agile coach will accompany you on the choice of the right approach and the right method, which will help to remove barriers.

On the difficulty of sequencing the work on the part of one of your collaborators, the work will be more on your side as a mentor of your team to accompany him or her. Although this long work requires an enormous capacity for hindsight on the part of the collaborator, who must change his way of doing things, it is your responsibility to place yourself as a benevolent coach and to find the right words to get your collaborator to put himself in the best possible frame of mind. There is always a risk of complete blockage, because it is still human, but it is up to you to persist and remain … agile. Act, Fail/Success, Iterate.

Why is it interesting to work with the Agile approach on Infrastructure?

In the end, you will reduce the risk of having to call into question the entire infrastructure put in place, and by extension, the major and very costly revamps that we are familiar with.

And you, do you think it is possible to work on Infrastructure topics using an Agile approach? Feel free to contact me to discuss it.

Gravatar de Francois Aichelbaum
Original post of Francois Aichelbaum.Votez pour ce billet sur Planet Libre.

Journal du hacker : Liens intéressants Journal du hacker semaine #8

lundi 24 février 2020 à 00:01

Pour la 8ème semaine de l'année 2020, voici 12 liens intéressants que vous avez peut-être ratés, relayés par le Journal du hacker, votre source d’informations pour le Logiciel Libre francophone !

Pour ne plus rater aucun article de la communauté francophone, voici :

De plus le site web du Journal du hacker est « adaptatif (responsive) ». N’hésitez pas à le consulter depuis votre smartphone ou votre tablette !

Le Journal du hacker fonctionne de manière collaborative, grâce à la participation de ses membres. Rejoignez-nous pour proposer vos contenus à partager avec la communauté du Logiciel Libre francophone et faire connaître vos projets !

Et vous ? Qu’avez-vous pensé de ces articles ? N’hésitez pas à réagir directement dans les commentaires de l’article sur le Journal du hacker :)

Gravatar de Journal du hacker
Original post of Journal du hacker.Votez pour ce billet sur Planet Libre.

Articles similaires

Full Circle Magazine FR : Le numéro 153 en français est sorti

samedi 22 février 2020 à 11:02

Bonjour,

Nous n'étions que deux ce mois-ci, car notre scribeur attitré se prélasse au soleil, le veinard ! N'empêche que, grâce surtout à d52fr, nous sortons ce numéro en français une semaine avant la publication du 154. Comme d'habitude, vous pouvez le visionner ou le télécharger sur notre page NUMÉROS ou le récupérer directement en cliquant sur la photo de couverture ci-dessous.

issue153fr.png

Pour ceux d'entre vous qui, comme moi, utilise Windows 10 aussi bien qu'Ubuntu, il y a deux articles qui le concerne (plus ou moins) :

Outre les bonnes choses habituelles (tutoriels, Q. ET R., etc.), il y a une opinion par Chris Binnie sur comment sécuriser vos conteneurs, qui vaut vraiment la peine d'être lu, une critique élogieuse de mtPaint, car il est plus simple et moins gourmand en ressources que GIMP, et une sur le jeu Stygian, basé sur le Mythe de Cthulhu, qui contient quelques bogues pour l'instant et qui coûte trop cher - vivement les soldes ! - mais les fans de Cthulhu seront comblés.

Nous vous en souhaitons très bonne lecture !

d52fr, qui a tout fait : scribeur, traducteur et relecteur et moi, AuntieE, traductrice et relectrice

Gravatar de Full Circle Magazine FR
Original post of Full Circle Magazine FR.Votez pour ce billet sur Planet Libre.

Articles similaires

Simon Vieille : Analyser efficacement un fichier CSV

jeudi 20 février 2020 à 15:17

Je travaille régulièrement sur des traitements de fichiers CSV. J'apprécie beaucoup ce format de fichier car il se génère et se lit facilement, qu'on soit informaticien ou non. J'ai d'ailleur écrit 2 librairies PHP pour lire, générer et valider des CSV : deblan/csv et deblan/csv-validator.

Lors des mes analyses, je me confronte parfois à des fichiers qui comportent beaucoup de colonnes et c'est un enfer d'utiliser un tableur pour visualiser et filtrer ces tableaux très larges.

Après quelques recherches, j'ai découvert deux projets qui se complètent à merveille ! Le premier est un outil qui transforme un fichier CSV en une base de données SQLite : csvs-to-sqlite. Le second génère une interface web (avec un serveur web intégré) pour faire des requêtes SQL sur une base de données SQLite : datasette.

Ce sont des outils écrits en Python qui s'installent et s'utilisent en une poignée de secondes.

$ sudo pip3 install csvs-to-sqlite datasette

Pour illustrer, je vais traiter le fichier CSV de la base officielle des codes postaux disponible sur data.gouv.fr.

$ wget -O code_postaux.csv https://datanova.legroupe.laposte.fr/explore/dataset/laposte_hexasmal/download/\\?format\\=csv\\&timezone\\=Europe/Berlin\\&use_labels_for_header\\=true
$ csvs-to-sqlite -s ";" codes_postaux.csv codes_postaux.db
$ datasette serve codes_postaux.db
Serve! files=('codes_postaux.db',) (immutables=()) on port 8001
INFO:     Started server process [8550]
INFO:     Waiting for application startup.
INFO:     Application startup complete.
INFO:     Uvicorn running on http://127.0.0.1:8001 (Press CTRL+C to quit)

On peut maintenant accéder à http://127.0.0.1:8001 depuis un navigateur et commencer à analyer et traiter les données :

datasette

Et le plus intéressant pour moi, c'est de pouvoir écrire des requêtes SQL affiner vraiment les données affichées :

datasette

Enfin, on peut générer un nouveau fichier CSV (ou JSON) avec les données filtrées.

Simple, rapide et efficace !

Gravatar de Simon Vieille
Original post of Simon Vieille.Votez pour ce billet sur Planet Libre.

Articles similaires

I'm richer than you! infinity loop