The problem with software for software teams is not that the big players are missing features. It's that every team wants something that is highly customized for them, and that leads to feature bloat.

Then people say, "Why doesn't Jira have ____?" In fact it does have that, but it's hard to find it in the massive UI that contains all those features.

I had this same issue testing Asana many years ago. I could tell that it could do what we wanted (or what anyone wanted), but I didn't want to spend 3 hours figuring out how to do it.

RE Jira - I've never seen a GANTT chart & critical path pulled out of it, though I imagine management 2 layers above me had that option. That said, my experience with Jira was somewhat painful, with a rather slow and annoyingly click-heavy UI. Also, IIRC, tasks in Jira can only have one level of subtasks, and I don't recall being able to form DAGs from them.

I imagine it's hard to prevent featuritis in this tool category, when you're addressing different hierarchy levels simultaneously. That said, there are many successful companies, like Trello, that thrive on delivering bare-bones, simplified tooling with perfected UX. I just wonder why there doesn't seem to be anything like that for a featureset that includes tasks forming DAGs, GANTT charts, computing critical path, and relative scheduling (e.g. "task Y takes starts after X and takes 5 work days").

JIRA is as system that is hilariously hobbled by two forces - JIRA admins and management, the latter often driving JIRA admins into making bad changes[1].

It has a very keyboard-capable UI that isn't highlighted enough (make it exempt from anything that might hijack keyboard, and hit . (dot) - it will bring you an autocompleting text box allowing you to do every action you can on given page).

The big issue however is that there's nothing really integrating the links system between JIRA items, and the basic task model is indeed rather limited. The links are fully free form (which can be a pain when someone defines a new copy of every standard link type, in german, without connection to localization system), but I have yet to see something that isn't the advanced query (i.e. the SQL-like queries you can type in various places in JIRA) that exposes it nicely and graphically.

That said, graph-based task representation is the norm on the higher end of _desktop_ project management tools, ones that often involve physical things to make, like MS Project or OmniPlan (and going back all the way to LisaProject, which apparently became very liked at NASA for exactly that graph based view).

[1] At one point, a friend of mine had enough and implemented exactly what the management asked for, without spending futile time trying to help them see the truth, or fixing it quietly by themselves. The results were glorious, as the whole project essentially ground to halt for a sprint or two, and could be a poster example of why people tell you to not mess and use provided standard workflows. Management never again asked for workflow changes after asking for reversal of their own design[2].

[2] Management, in their infinite wisdom, designed a one-way-only workflow, where a task could proceed a workflow a bit like this: Backlog -> TODO -> Development -> QA -> Done. Notice lack of any way to move a stage back, and there was no way to force change a task to different stage. Tracking real status of tasks became impossible as everything was left at "Development" because if you found anything to fix after it left QA... well, no way to change the state without hacking the database, which might have not gone well with Atlassian Cloud admins ;)

> and hit . (dot) - it will bring you an autocompleting text box allowing you to do every action you can on given page

How I wish I knew that at my {previous}^2-job, where we were heavy JIRA users. I remember getting so angry off at the frontend performance that I actually (somehow) got minimal API access and ended up timesheeting from a Lisp REPL.

I thought about my previous comment through the night and realized that, at some point, I should have had found a reference manual for JIRA and read it end-to-end (reading manuals cover-to-cover is a superpower, but that realization only truly hit me around the end of said job). Also, I now know enough to articulate what was my biggest discomfort with JIRA (beyond abysmal performance): lack of clear mental model of what it does, and what is being represented inside. This came about not just because the product is large, but also because JIRA admins - we (developers) had access restricted only to a tiny portion of the product, so I couldn't just poke around (my usual way of grokking software) and build a clear mental picture. Next time in such situation, I'll just go straight into "read docs cover to cover" mode.

> That said, graph-based task representation is the norm on the higher end of _desktop_ project management tools, ones that often involve physical things to make, like MS Project or OmniPlan (and going back all the way to LisaProject, which apparently became very liked at NASA for exactly that graph based view).

I suspected so! I saw some of these capabilities during my brief exposure to the "corporate pointy-haired MBA" project management tools during my time at the university. But I get the feeling that it's another place where Agile culture is throwing the baby out with the bathwater.

You can then make a mental model that ultimately JIRA has project, which have a "table" of issues/items, with various attributes each - plus simple workflow engine - everything else is just built around it, whether kanban boards, SCRUM, various reporting tools (very powerful), or random plugins (integration with SCM, time tracking, etc. etc)

Really cannot wait for Google Workspace to offer real task/project management features.

It's not open source though

It's a visual DAG that auto schedules out the plan given the constraints. It's not for everybody - but for highly visual people it seems to resonate really strongly.

If you have any questions please email me on phil@gameplan.global

I use it all the time, I looove it for early project scoping.

This is taken from an old project proposal of mine, done in OmniPlan 3, and shows causal dependencies between components more than time-based scheduling (like in GANTT chart)

MS Project and many other task tools recognize that, considering all of start-to-start, start-to-finish, finish-to-finish, and finish-to-start dependencies (all of which can be simultaneous as well as happen-first dependencies), tasks can form directed graphs (not restricted to acyclic).

And many of those tools also have GANTT charts built in.

For personal use, as in general tasks not necessarily code for any particular project, so I haven't really envisioned any such integration.

I just get bogged down/stressed thinking of numerous things I want/need to do, there's some interdependence, a few may occur to me at a time or be particularly weighing on me on any given day, and I'd like a way to graph the dependence, actually view it as a graph, annotate some tasks with their priorities (but not all, some might be 'don't care'), and then find out which is the next thing I need to do.

Since the next thing isn't necessarily the highest priority thing, as annotated (or stressed about), it's the first thing on the path (dependency chain) to the highest priority thing. Attempting to see that clearly as a mental picture is what causes me most stress, I think.

No limit on task nesting, gantt charts built in, git integration, and very customizable.

I'll just say that redmine DB schema is pretty simple, and metabase works well with it if you need more complicated reports :)

Another thing to keep in mind is that there's a distinction between a system having cyclic dependencies (I need A & B working because they communicate with each other), and the task ordering. During development you can end up with a partially correct system. If A sends to B, and B sends to A but I only have one person working on both parts, then it's possible I'll end up in a state where A sends to B but not the reverse. That's fine during development. Yes, the system is "wrong" per the design or spec. But it's a partial system (like a half-built bridge). Your tasks will appear to have a cycle, but in reality you will break it by selecting one portion to complete first so in the end it remains (based on the chosen path) a DAG.

That produces subtasks without cycles, giving you a way forward.

Assuming you still want to maintain shared information about the tasks which contain subtasks, those tasks still have a dependency cycle between them though, which is correct because tasks are often like that - they need to move forward together.

And that's why I'm asking for a tool that lets me model projects as Directed Acyclic Graphs of tasks. Because all I seem to be getting in popular tools is a flat list with some categorizing, or a single level of subtasks at best. I want to be able to express directly that this subtask has sub-subtasks, and that this sub-subtask depends on completion of a subtask from some other task (without making it a dependency between root tasks).

Real dependencies can be not-before rather than strictly-after, which admits cycles without any problems with the flow of time.