Participating a little

19 views
Skip to first unread message

Yannick Warnier

unread,
Dec 13, 2009, 6:01:43 PM12/13/09
to js-mindmap
Hi Kenneth,

First of all, thank you for your work. It's a great, innovative little
library you've written there and it gave me great ideas for visual
classification of items inside a GPL software I'm working on.

Now, I have a few questions and a few suggestions.
I'd like to know what's the direction js-mindmap is going and if you
plan on maintaining/improving the library in the near future, and how.
It's important for me to get an idea of whether I should wait, get
involved, or look for something more adapted to my needs.

I'd also like to know why you moved from version 1.0 to version 2.0
and if the loss of features (like auto-hiding the elements at more
than one step and painting the links red for current-to-father
relations) is a temporary state, or if it is meant to stay like this
for a while.

I've been playing around a little with the library and it seems to me
that the move to jQuery might have been done a bit too hastily, as
there is a lot of mixed code between pure JavaScript and jQuery
objectsn, which makes it a bit difficult to handle.

I have a modified version of the js-mindmap.js which, mostly, contains
additional documentation about what the code does (which I bet will
help get more people involved).

I have also tried to figure out a better way to represent elements and
their father-current or current-children relationships, but so far I
haven't made much progress here.

My objective is to have a tool that goes a bit more into the mindmap
area than the sitemap philosophy you are using as an example. So
basically, I'm not sure the <ul><li> structure will suit my needs. I'm
currently trying to think of an acceptable HTML structure, but I'm
pretty sure I'm going to have to use specific ID's for each element,
as well as a separate element for each line between elements.
This way, I can draw as many elements as I need and have them relate
through other elements of type "link" which, in turn, can all specify
a different type of relation.
This might imply going out a little of the HTML idea, and getting more
into an XML relationship (maybe re-using the FreeMind XML format to
save efforts).

So I'd be pretty interested in knowing all about the intentions of the
project, in order to decide between a pure branch/fork or joining
efforts with you. My jQuery level is beginner, and I have generally
very little free development time, but I have ideas and I need to get
them realized at some point.

In any case, thanks for showing the way.

kenneth

unread,
Dec 14, 2009, 5:23:05 AM12/14/09
to js-mindmap
>>I have generally very little free development time

My problem too. It's been a busy year. I hope to turn my attentions
back to this in the not-too-distant future.

I like your suggestion of the XML basis. The HTML list is a good
approach, but begins to fail rapidly. There is simply no way to have
multiple parents this way.

I think the plan for this project is experimentation rather than
product-building, so if this helps you get started on a mindmap
project, then I think that's great. It'd be great if you can share
your results back on this list, for others to follow, and if you have
fixes to contribute back, we can sort that too.

Cheers
K

Yannick Warnier

unread,
Dec 16, 2009, 4:32:12 PM12/16/09
to js-mi...@googlegroups.com
OK, thanks for your answer. One version of the js-mindmap (version 1,
I think), is available here http://brookellingwood.com/mindmap from
someone who also went for the XML path.

As I'm trying to figure out what I want to do, I realize more and more
that the classical mindmap schema is not what I need. I need something
like a "thesaurus", where nodes can be linked to others through a
"synonym" relationship, so I would have three types of relations:
child to father, father to children *and* synonym to synonym.

My biggest problem here is the maths to draw the links between
elements that have these 3 links. Do you see this third type of link
like something that should be handled easily by the current library?
It introduce a kind of risk of "loop" in the process. For example:

"Car companies" is father of

Ford Mercedes Porsche Ferrari Lotus Tesla BMW VW

but I want to say that VW is the same company as Porsche.

This means that I will need to show links between "Car companies" and
all its children, but I will also need to show a link between VW and
Porsche.

Any thought on this?

Because time is scarce for both, I want to avoid going in one
direction and then going back because I didn't think thoroughly
enough.

Yannick



2009/12/14 kenneth <kenneth...@googlemail.com>:
> --
>
> You received this message because you are subscribed to the Google Groups "js-mindmap" group.
> To post to this group, send email to js-mi...@googlegroups.com.
> To unsubscribe from this group, send email to js-mindmap+...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/js-mindmap?hl=en.
>
>
>

kenneth

unread,
Dec 21, 2009, 12:23:52 PM12/21/09
to js-mindmap
Hmm. The js-mindmap certainly would support links like that. Since
it works on nodes and links, you just define more links.
It's not set up to do so out of the box, because HTML lists are
hierarchical - no cross-links.

So we need a new XML definition. I've been looking at the FreeMind
docs. Couldn't find a good definition of the markup - do you have
one?
Otherwise, I guess we can just extend what we have.
<node>
<subnode></subnode>
</node>
And then throw in additional nodes or links at the bottom:
<link node="" node="" type=""/>
<node>Some other node that doesn't logically fit in the hierarchy</
node>

Semantically, the node elements could be LI elements. But I can't
think of a semantic equivalent of a two-way link between elements on a
page. I keep thinking about anchor tags, but they don't really fit
the use-case.

K

On Dec 16, 9:32 pm, Yannick Warnier <ywarn...@gmail.com> wrote:
> OK, thanks for your answer. One version of the js-mindmap (version 1,
> I think), is available herehttp://brookellingwood.com/mindmapfrom
> someone who also went for the XML path.
>
> As I'm trying to figure out what I want to do, I realize more and more
> that the classical mindmap schema is not what I need. I need something
> like a "thesaurus", where nodes can be linked to others through a
> "synonym" relationship, so I would have three types of relations:
> child to father, father to children *and* synonym to synonym.
>
> My biggest problem here is the maths to draw the links between
> elements that have these 3 links. Do you see this third type of link
> like something that should be handled easily by the current library?
> It introduce a kind of risk of "loop" in the process. For example:
>
>            "Car companies" is father of
>
> Ford   Mercedes   Porsche   Ferrari   Lotus   Tesla    BMW    VW
>
> but I want to say that VW is the same company as Porsche.
>
> This means that I will need to show links between "Car companies" and
> all its children, but I will also need to show a link between VW and
> Porsche.
>
> Any thought on this?
>
> Because time is scarce for both, I want to avoid going in one
> direction and then going back because I didn't think thoroughly
> enough.
>
> Yannick
>

> 2009/12/14 kenneth <kenneth.kuf...@googlemail.com>:

Yannick Warnier

unread,
Jan 13, 2010, 9:27:46 PM1/13/10
to js-mi...@googlegroups.com
I don't have any good freemind schema and I'm not sure there is. I've
dropped that path.

I've been trying to think very hard of a way to do this well, but as
you say... it doesn't fit really well in the HTML DOM apart by maybe
using double lists or so (defining elements twice).
For my application, I have found a way though: I will just split the
screen into 4 divs: the current element, its parents, its children and
then its synonyms, and I'll manage the link through PHP and a database
backend (which is where my data comes from in the first place anyway).

So that wil probably not be extending js-mindmap in any way. Sorry
about that, looks like a vain effort for now.

Yannick

2009/12/21 kenneth <kenneth...@googlemail.com>:

Reply all
Reply to author
Forward
0 new messages