Site in costruction

Exobrain

Home of inorganic intellectual chaos and unpolished content

Kanban Software Review

Introduction

This log is the result of an exploratory study conducted by @qub1750ul and @f_dinucci evaluating "cloud-ready" Kanban software for coordination purposes in GARRLab.

The views, thoughts, and opinions expressed in this article belong solely to the authors, and do not necessarily reflect the official policy or position of any other organization or individual. While every caution has been taken to provide readers with the most accurate information and honest analysis, please use your discretion before taking any decisions based on the contents of this article. All information in the following is provided "as is", with no guarantee of completeness, accuracy or currentness.

Requirements

Planka

Reviewed on 2022-07-24

Planka is a user-friendly Kanban built using React.js.

At the time of writing the software is still in beta stage:

  • lacks basic features such as username search and sort. (GH issue)

  • some configurations are only half-suported.

  • UI is sleek, but problematic: not tested with more than a few users, UI bugs appear as soon as user number increases. (GH issue)

  • Updates are quite complicated and require badly tested (at all ?) procedures.

  • Update paths for breaking changes aren’t managed appropriately, and often impraticable. (GH issue)

  • Lacks any form of external authentication.

Judgement

Cool UX, but doesn’t meet basic requirements for our use case. Notably SSO.

Wekan

Reviewed on 2022-07-24

Wekan is an enterprise-grade Kanban built using MeteorJS.

  • The deployment procedure is…​ weirdly over-documented, with a LOT of useless repetition, at the point that it is difficult to read.

  • The project website and info on the software in general is VERY chaotic and all over the place.

  • Dubious approaches in maintaining auth features, and questionable methods to approach standards and interop quirks implementation

  • Disregard of fundamental security best practices:

  • UX is enterprise grade, in the sense that is an "enterprise grade mess"™:

    • UI is unpolished in a lot of places, it may appear broken when in reality it’s not

    • UI is confusing to approach in general

    • A lot of UI options aren’t documented ( even tooltips are missing ), and is not inutitive to understand how they work.

    • Head developer refuses to write user documentation, as it is deemed not important because "the software has proven monkey-proof". (See discussion on GitHub)

  • Has a variuous external auth options, but management of these is kinda cumbersome.

  • Disparity in features among builds for different platforms, even trough it is a Node.js webapp built using a JS framework whose main feature is to "unify server and client environments" (???).

  • Various usability problems dating back at least to 2020, still unsolved as of today and unlikely to be solved soon.

Judgement

Has interesting features. UX confusing for simple usage and a nightmare for complex workflows, also a security nightmare. Cumbersome documentation and dubious development practices.

Taiga

Reviewed on 2022-10-22. Missing some details from @qub1750ul

Taiga is a project management tool for Agile teams.

  • Deployment and Docs: @qub1750ul up to you

  • Mainly revolves around Scrum methods. It is possible to use it as a Kanban, but still the workflow is hevily influenced by Scrum, with user stories instead of tasks, Epics that should group feedback from the user stories and a points system to determine more important tasks (points for UI/Design/Front/Backend). Oh, the point system can’t be disabled. And the labels can’t be changed (eg. Ops instead of UI)

  • Rewards&Badges for users, based on eg. closed issues (Mr. Wolf) or who-knows-what (Iocaine drinker). Do we seriously need it?

  • Useless features in our use-case, such as issues, that are better managed on GitLab (due to linking to MR etc); or the whole Scrum section. We are not in a enterprise environment, sprints/due dates/customer requirements have no meaning. Or the whole part about interaction with external users, as this should be an internal-use-only tool.

  • Videochat integration: practically non existent, it is a link to an external service such as Jitsi. Btw, it lacks a text chat.

  • Wiki: nice to have, but we’re using GitOps and Hugo for the documentation. And, by the way, it is really too basic to be a proper documentation wiki

Judgement

Lots of features are not fit for our use case, it would just become a more complicated and resource-hungry version of Planka ~fdinucci

Mattermost Boards

Currently in review

Mattermost boards is a kanban board platform available both as a plugin for the Mattermost platform, or as a standalone single-user application ( Focalboard ).

  • It suffers from dreadful SSO Tax: the free version has very limited authentication options. In particular we found that OIDC auth is artificially limited to GitLab instances. Fortunately for us self-hosted instances are supported, and our GitLab supports the RP role.

  • UX is polished, intuitive and overall user-friendly.

  • Boards are very flexible: for each board a multitude of views can be created, allowing to have different customized perspectives on the same data.

  • It allows for flexible metadata management: each card can be arbitrarily enriched with strongly-typed structured fields.

  • Card metadata is reactive: it is smartly and automatically updated when it makes sense.

Judgement

TBW

Final considerations

TBW