[prev in list] [next in list] [prev in thread] [next in thread] 

List:       graphviz-interest
Subject:    [graphviz-interest] Dynagraph 1.0(.01) released
From:       gordon () woodhull ! com (Gordon Woodhull)
Date:       2006-01-25 12:03:00
Message-ID: 39C3CE85-8095-46E5-AECE-E318283D9138 () woodhull ! com
[Download RAW message or body]

Hi Axel,

You are so right.  I always wanted an object-oriented interface for  
two reasons:
1) I just think that way.
2) It allows for stronger typing of attributes, and you don't have to  
parse e.g. edge bezier strings.

However, the functional 'gv' API (based on lib(a)graph) is far more  
compatible, and I already have the string attributes from the  
incrface language interface.  So it should be much easier to  
implement, once I have time.

Thanks for a push in the right direction,
Gordon

On Jan 20, 2006, at 4:46 AM, Axel Belinfante wrote:

> I don't have much to offer, except the following (obvious?) idea:
> _if_ the current tcldot interface is rich enough to be usable
> as (the basis of) an interface for dynagraph, then I'd think
> twice before inventing a new API for tcl.
> The tcldot API already has commands to manipulate
> (extend, shrink) the graph, AFAIK.
> Of course it might be necessary to add support for dynagraph
> features not present in (tcl)dot.
>
> As an 'end user' who only uses part (subset) of the tcldot API
> I would not be unhappy if this subset also would be a subset
> of a tcldynagraph API - i.e. I could switch between tcldot and
> tcldynagraph without changing my 'application'.
>
> I guess similar reasoning might apply to other language bindings
> (interfaces) to dot(like) functionality.
>
> I have (still) not yet looked at dynagraph, so this ideas might
> be completely infeasable...
>
> Axel.
>
>
> Gordon responded:
>> Yes, Dynagraph could definitely use some help in bindings to other
>> languages (currently just C++, COM, and a dotlike "incrface"
>> language).  Unclear what the correct API and glue is for the task.
>> Welcoming any ideas and/or help over on dynagraph.org.
>
>
>> On Jan 16, 2006, at 8:25 AM, Axel Belinfante wrote:
>>
>>>> You can for instance input one graph in dot format and then "segue"
>>>> to another graph; Dynagraph will determine what was inserted,
>>>> modified, and deleted, and try to make the minimal changes to the
>>>> first layout to display the next.  (Or of course you can specify  
>>>> the
>>>> insertions, modifications, and deletions explicitly.)
>>>>
>>>> Dynagraph supports both dot-like directed/hierarchical graph  
>>>> drawing
>>>> and neato-like undirected/springy graph drawing.  It calculates
>>>> layouts of up to 100 nodes & edges in less than a second.
>>>
>>> Currently I'm using tcldot to animate the construction (and
>>> exploration)
>>> of the state space of a labelled transition system, using tcldot's
>>> features to add nodes and edges to an existing graph.
>>> Does (will?) dynagraph have similar tcl bindings?
>>>
>>> Just curious.
>>> (sorry for the noise if I should have been able to find this easily
>>> myself)
>>> Axel.

[prev in list] [next in list] [prev in thread] [next in thread]