Can anyone explain to me why most mind map apps are of tree structures, but not graph structure?

In my personal process of thinking most ideas pop up as nodes then nodes are interconnected as a graph.

I tried to use one of these tree mindmap apps, some ideas morphs into another leaf on another subtree but I can not connect them visually, this irritates me a lot.

An acyclic graph (or a tree structure) enables certain automatic layout algorithms. These algorithms are predictable by the user and are less distracting -- the key goal of any mindmap software is to not distract.

There's probably a semantic element of a hierarchy of ideas, too, but I am only equipped to comment on the math and not the psychology. :)

Yes a directed acyclic graph would help a lot.

But there isn't any mind map apps support that. (Not that I know of)

I now use yEd and OmniGraffle from time to time

I use OmniGraffle when a mac is handy; thanks for the reference to yEd, I hadn't seen that one before.

Good for yEd find.

So if we were going to take the LaTeX approach, we basically write our mind maps as .g files and then lay them out using Graphviz's dotty? That could actually work (OmniGraffle, BTW, uses Graphviz I think to layout DAGs).

If you're on a mac try "Scrapple". I had the same issue and Scrapple is incredible.

I think it's "Scapple". Thanks for the tip, it looks good!

Yep, Scrapple is a fairly gross pork product popular in Philly haha. Typo.

That's pretty cool. After playing around with this some, I'm glad that I won't just get to use it, but that I'll also get the chance to learn something from the code. I don't believe I've ever seen such nicely routed auto-updating paths before in Javascript.

Thanks. We still need to clean up the code a bit, and should be publishing more docs on how the thing is structured this week, so keep your eye on our github repo.

Great work! A suggestion: to make the text searchable using Ctrl-f, please render the maps using SVG instead of canvas. For a big mind map this is indispensable.

is search the only thing that would be better with SVG?

we opted for canvas as it gives us better visualisation capabilities in the future. we could easily provide searching with canvas by overriding ctrl+f.

I haven't tested this but I think search engines should automatically be able to index SVGs, thus making the public maps Googlable.

In general I like SVG because it doesn't try to hide the data it's displaying, unlike flash/canvas. Although I now remember seeing a hybrid div/canvas implementation which buys you best of both worlds albeit at some added complexity.

Good thinking - we already planned an svg spike (probably using raphaeljs) :)

Hats off for open-sourcing this, especially under the MIT license, could make a great add-on for an Saas app!

Thank you for this. I see this as a perfect companion to Freemind. Sometimes I want to jot stuff down, and firing up freemind is too cumbersome, or I'm on my laptop and haven't sync'd recently. I've been using Evernote for this purpose, and it's just not quite as graceful as a mindmap.

For getting ideas out quickly, and sharing mindmaps without installing software or paying for an one of the (many) web-based services, this looks pretty ideal.

Also, +1 for folding.

Yes, that's why we built it as well. The basic use case is knocking up maps quickly during a meeting. Please vote for folding on the site if it's important, we aim to invest as much as we can on working on the top voted feature group.

I don't see this as being useful unless it share's FreeMind's 'strong emphasis on folding'. Otherwise it's basically like a text editor that doesn't let you create documents longer than one page.

I also probably wouldn't consider using it until at a minimum the data is exportable and the data format is future proof.

export to freemind is done, will go live tomorrow. folding is also in the plan. you can make it go live earlier by - voting directly on http://www.mindmup.com/#vote - or forking on github and submitting a patch :)

Great work on keeping the interface concise, consistent and keyboard friendly. Overcomplex UI has kept me from using mindmap tools in the past, since the tool becomes a distraction.

A "yank" and "place" for nodes would be a nice feature, as I still have to reach for the mouse when rearranging. "Yank" should mark the currently "yanked" node with an different style, and "Place" should take the "yanked" node and set its parent to the currently selected node.

This has many features similar to MindNode, but I miss the ability to drag and drop a subtree. Out of all the online mind mapping apps I've seen this one is the best.

You should be able to do this (you can't however drop a subtree to a node that's in the subtree itself)

Is there a plan to make it work offline _without_ needing S3? i.e maps are saved to the local hard disk? Great work though - looks awesome.

we could probably make it work offline as a chrome app, not sure if offline work with all browsers is a viable option. would chrome app be good enough for you?

Can't answer for OP, but for this new user, a Chrome app would be dandy.

Absolutely - Chrome app would be perfect.

Wow! Simple but effective. I can't test multi-user right now, but assuming that works I'd love to see this in Google Apps.

We need to think beyond mind maps. Not everyone is able to understand it.