I would like to know what you think of this use case:
When I have (mostly business) meetings I normally take my notes on the
laptop. But until now I haven't found the right to tool that matches my
needs. The main problem is, that I write my notes down in a text editor.
Later on I have to go over it and generate tasks, mails, documentation
etc. out of it. The main advantage compared with hand written notes is
that I can copy/paste the text. Another problem is that normally I have
meetings that concern mutliple projects - so where should I best file my
minutes?
So my use case is: Take my minutes in a topic map like I sometimes do it
with a mind map tool. The topics have functionalities like "send mail
to", "create a task in planning tool", "create documentation in project
folder", "create page in wiki", etc. etc.
Now, what do you think of that proposal? At the moment it is just an
idea, but I would love to work it out properly with anybody who's
interested in it.
Best
Urs
Thanks for your input. Generally spoken I completly agree with you. But
I don't see my proposed use case competing with the software tools you
named.
If I gave you the impression that I would like to implement all the
functionality of a mail program, a task manager, a document management
system etc. etc. in DeepaMehta, I feel sorry. What I really meant is
that I would like to take my minutes inside DeepaMehta. I think the tool
is meant for that. And now, instead of distributing all the information
afterwards with copy/paste to all the places (which means mostly
applications) where I need it, I would like much more to have the
functionality to feed the APIs of these different applications (Outlook,
Bugzilla, Alfresco etc.) instead.
I don't think that it is so much work to code the desired functionality.
Much more it is important to have a good design, so it can be used in as
many cases as possible.
Best
Urs
I would love to work this out together and help with implementation.
Unfortunately for the time being i'm occupied (with the Freifunk use case).
However, I like to throw in some consideratons here.
1) Modularization.
The features you describe could be realized as separate DM modules=plugins. The user would only install the plugins he likes to use. This way the UI doesn't get cluttered.
So, what would make up reasonable plugins here?
E.g. there could be a "deepameht3-wiki" plugin which would connect DM to existing Wiki's. The plugin would attach a "Create in Wiki" command to every DM Note topic (or every DM topic? or a topic map?). This command would create a Wiki page from the topic's content.
Accordingly there could be e.g. deepamehta3-email and deepamehta3-tasks plugins.
2) Customization
The user would need a mechanism inside DM to configure the external applications. There she would e.g. choose the Wiki software in use, set up the Wiki's Base URL, and possibly the user credentials.
3) GUI
The extra commands could easily be added e.g. to the topic's context menu.
In general, the existing DM plugin framework makes it easy to add commands to the context menus (topic's, relation's, canvas's), the detail panel, or to the toolbar.
Customization could be realized through a preferences dialog, with tabs for the various function areas (not yet existing). Plugins could add tabs there to provide their particular settings.
Urs, if you like to start such plugin projects at your own, I'll provide you with required DM API infos and possibly with little core extensions.
Cheers,
Jörg
I don't think that it is so much work to code the desired functionality.
Much more it is important to have a good design, so it can be used in as
many cases as possible.
-- -- --- Maintain your head and share it --- -- - http://blog.deepamehta.de - skype: einseinsnull - jabber: mre at newthinking.de - mail: mre at deepamehta.de - fax: 01803-663388-72229 - tel: 030-42800182 |
> This is not the only thing you can do with this "Bulk Creator" by now it allows you to import various source, like HTML Tables or simple .CSV into topic/datafield structured informations. Currently i have no plans to develop this any further.
On Nov 1, 2010, at 12:19, x28 wrote:
> Furthermore, the place where the imported icons are inserted, must be
> more sophisticated, because currently, they keep popping up in the
> center, cluttering your carefully positioned previous ones.
Hi Malte!
You could easily realize x28's suggestion by using the canvas's "grid positioning" facility.
What it does:
- finds a free space on the canvas (so, subsequently placed topics will not overlap with anything)
- position topics in a grid-like manner (by respecting the canvas's viewport width)
- scrolling the canvas to bring the grid's upper/left position into view (useful for large grids)
Additionaly you could select the first imported topic prgrammatically (to provide the user a visual clue what was imported).
(All in all it would be the same behavoir as shown by the "File Manager" and the "Folder Canvas" plugins.)
You would just need to add this code:
dm3c.canvas.start_grid_positioning()
<your loop that creates the topics> {
...
dm3c.add_topic_to_canvas(topic, <is first topic> ? "show" : "none")
}
dm3c.canvas.stop_grid_positioning()
Selecting the first topic is achieved by passing "show" to dm3c.add_topic_to_canvas() for only the *first* topic, and "none" for all other ones.
Cheers,
Jörg
> I would love to show to them my example thread
> http://www.flickr.com/photos/37794987@N00/2854506800/ in DeepaMehta,
> but after TWO years I still CAN'T (without urging them to install or
> register which they don't want, http://ple.elg.ca/course/moodle/mod/forum/discuss.php?d=458
Perhaps the DeepaMehta 2 "Demo Mode" would be helpful for you.
It exists since DeepaMehta 2.0a8 (January 2001) but is not well documented.
The "Interactive Demos" page is build on it (try out the "History of Human-Computer Interface" demo):
www.deepamehta.de/install/client/demos.html
About the DeepaMehta 2 Demo Mode:
- allows you to make a DM2 topicmap accessible to everyone
- for the demo user no software installation is required (just a Java-enabled webbrowser)
- no registration is required
- the demo user can:
- open a topicmap by visiting a webpage and pressing a button
- browse the topicmap
- arrange the topic geometry
- revealing (still invisible) related topics (!)
- the topicmap is read-only:
- the demo user can't create or edit anything
- the topic geometry is not saved
- once the demo user registers she might be able to make changes to the demo topicmap
The purpose of the DeepaMehta 2 Demo Mode is to provide users a quick way to experience DeepaMehta interactively.
Every user who is registered at www.deepamehta.de can setup a demo topicmap for public access. The demo topicmap can be embedded into your own website/blog by the means of a HTML snippet.
To setup a demo topicmap for public access:
1) Login to DM2 at www.deepamehta.de
2) Create or import a topicmap
3) Publish that topicmap to any (shared) workspace
4) Embedd this HTML snippet in your webpage:
<applet archive="DeepaMehtaClient.jar" code="de.deepamehta.client.DeepaMehtaClient" codebase="http://www.deepamehta.de/install/client/" width="150" height="30">
<param name="DEMO_MAP" value="t-1147">
<param name="PORT" value="7557">
<param name="BUTTON_LABEL" value="Open Topicmap">
<param name="BACKGROUND_COLOR" value="#ffffff">
</applet>
This will create a litte "Open Topicmap" (or whatever you label it) button on your webpage.
The only tricky thing here is the DEMO_MAP parameter. There you have to state the internal ID of your demo topicmap (t-1147 in this example). This ID is not exposed to the user interface but there are ways to determine it anyway (let me know if you need support at this point).
5) Distribute the URL of your webpage to interested parties and instruct them to press the "Open Topicmap" button.
So, this would be roughly the state of affairs when you want to provide any internet user instant access to your topicmaps.
In DeepaMehta 3, making topicmaps publically accessible will become less cumbersome. As soon as the DM3 "Access Control" plugin is ready (within this year) there will be a server-hosted DM3 installation (like DM2) and an opportunity to directly address any topicmap by the means of an URL.
Cheers,
Jörg
On Nov 1, 2010, at 11:25, Malte Reißig wrote:
> This is not the only thing you can do with this "Bulk Creator" by now it allows you to import various source, like HTML Tables or simple .CSV into topic/datafield structured informations. Currently i have no plans to develop this any further.
would it also be possible to use this for the manual
creation of a topic via the menu "+Create" command?
Right now v0.4.2 it happens rather often that the new topic is
on top of another one and draging will first move the "wrong"
topic.
Thanks,
Torsten
Am 03.11.2010 15:13, schrieb J�rg Richter:
> Hi Malte!
> You could easily realize x28's suggestion by using the canvas's "grid positioning" facility.
> What it does:
> - finds a free space on the canvas (so, subsequently placed topics will not overlap with anything)
> - position topics in a grid-like manner (by respecting the canvas's viewport width)
> - scrolling the canvas to bring the grid's upper/left position into view (useful for large grids)
>
> Cheers,
> J�rg
>
--
Torsten Ziegler
tor...@ziegi.de
To solve the overlapping problem 2 solutions come to my mind:
1) An algorithm that actually looks for a free position for placing a single topic *within* the current viewport (and would resort to scrolling only if the current viewport is filled-up).
2) An alternate gesture for creating a topic: right-click on the canvas and choose "Create Note" (or whatever type is selected in the upper toolbar) from the canvas's context-menu (similar to DM2). Thus, the user would have full control about the initial topic position *and* must not fumble with submenus (like in DM2).
The upper toolbar's "Create" button could remain as it is. So, there would be 2 alternate ways to do the same thing (the Create button and the context-menu), but with a reason: the Create button would still have the advantage of *being visible* as a lot of users would not intuitively right-click an empty spot (user feedback showed that).
What do you think?
I could realize the changes quickly and provide a new release next week.
Cheers,
Jörg
On Nov 3, 2010, at 15:32, Torsten Ziegler wrote:
> Hi Jörg,
>> Jörg
>>
>
>
> --
> Torsten Ziegler
> tor...@ziegi.de
now your demo map is set up.
Paste this HTML snippet on your website or blog:
<applet archive="DeepaMehtaClient.jar" code="de.deepamehta.client.DeepaMehtaClient" codebase="http://www.deepamehta.de/install/client/" width="150" height="30">
<param name="DEMO_MAP" value="t-91204">
<param name="PORT" value="7557">
<param name="BUTTON_LABEL" value="Open Topicmap">
<param name="BACKGROUND_COLOR" value="#ffffff">
</applet>
You can choose any button label you like.
BTW: the server-hosted DM2 installation is repaired now. Importing topicmaps works again.
Thanks for the hint!
I'm wondering how the demo map works on Windows.
On a mac there are crucial problems, e.g. the actual topic contents do not appear! and the demo user can edit content! I guess this is because of incompatibility of the old DM2 with current Java versions.
However, I figured out if you use the *signed* applet things run much better and the mentioned problems do not occur (on a mac).
To use the signed applet, just change
archive="DeepaMehtaClient.jar"
to
archive="sDeepaMehtaClient.jar"
IMPORTANT: after changing this restart the browser.
However, the drawback with the signed appplet is, the user must do one more click to say she trusts the applet.
One more hint for effective use of the demo mode: it is possible to set up the demo map in a way it contains only some starter topics and let it to the user to reveal more content exploratively (by "What's related"). Thus, the user would not get clobbered with a huge map, and the browsing/exploring experience would become more personal and possibly more adequate cognition-wise. The idea is to start with a small clear map and allow the user personal exploration.
To do this, just login, open the demo map (its in the "DeepaMehta Community" workspace), hide topics, re-arrange geometry, and finally publish it again (to the "DeepaMehta Community" workspace). No changes are required on the applet-tag.
Cheers,
Jörg
this is a good point, that scrolling ist not what we want on
creating an new topic.
I think number 1 would be pretty good for the beginning
maybe with the rule to find the next free spot near the last
active topic and if there is plenty of space to favour the
right side and a little bit lower than the last active topic
This is not to urgent so i don`t think ist is necesary to
provide a new relase only for this.
Number 2 would dig into the whole paradigm of the user
interface and i think it would be good to share and discuss
some thoughts on this before you do too much work there.
Some of my visions are:
- clicking with the mouse to create new topics
- creating of associations without the need of the drop down menu (like:
http://morrisonpitt.com/jsPlumb/html/jquery/demo.html)
- creating topics by just typing on the canvas
- navigation in the canvas by using arrow keys
Lately I used FreeMind a lot again because it is so much
faster to develop thoughts. Using the mindmap 10 minutes
the same process would would take at least an hour in
deepamehta. But it is missing the free positioning.
So I would also like to look a new topic type that
allows to create related or "sub" topics just by typing
on the active topic in the canvas.
And also it should be possible to show properties
that are normally only visible in the detail pane
in a mind map style fashion around the topic in the canvas.
... I am still thinking about this.
Yours,
Torsten
Am 06.11.2010 15:38, schrieb J�rg Richter:
> From my point of view the "grid positioning" facility would not be a perfect fit for the problem you describe. Grid positioning is meant to place an *arbitrary* number of topics. It always starts positioning *below* the topicmap's bottom-most topic and sets the viewport accordingly (by automatic scrolling). So, when you're working in an upper region of a large map and create a topic your viewport would be lost.
>
> To solve the overlapping problem 2 solutions come to my mind:
>
> 1) An algorithm that actually looks for a free position for placing a single topic *within* the current viewport (and would resort to scrolling only if the current viewport is filled-up).
>
> 2) An alternate gesture for creating a topic: right-click on the canvas and choose "Create Note" (or whatever type is selected in the upper toolbar) from the canvas's context-menu (similar to DM2). Thus, the user would have full control about the initial topic position *and* must not fumble with submenus (like in DM2).
> The upper toolbar's "Create" button could remain as it is. So, there would be 2 alternate ways to do the same thing (the Create button and the context-menu), but with a reason: the Create button would still have the advantage of *being visible* as a lot of users would not intuitively right-click an empty spot (user feedback showed that).
>
> What do you think?
> I could realize the changes quickly and provide a new release next week.
>
> Cheers,
> J�rg
>
>
> On Nov 3, 2010, at 15:32, Torsten Ziegler wrote:
>
--
Torsten Ziegler
tor...@ziegi.de
> this is a good point, that scrolling ist not what we want on
> creating an new topic.
>
> I think number 1 would be pretty good for the beginning
> maybe with the rule to find the next free spot near the last
> active topic and if there is plenty of space to favour the
> right side and a little bit lower than the last active topic
Let DM position a new topic *in the near* of the active topic is a good idea.
And when there is no active topic the new topic could be placed near to the Create button (to visually group cause and effect).
> This is not to urgent so i don`t think ist is necesary to
> provide a new relase only for this.
>
> Number 2 would dig into the whole paradigm of the user
> interface and i think it would be good to share and discuss
> some thoughts on this before you do too much work there.
Why do you think approach number 2 would undemine the DM user interface paradigm?
We have context-menus already (and the existing approach to create a topic would persist).
Both approaches would directly address the problem you described in the first place (better placement for new topics). Both are sufficiently clear and could be implemented quite easily (approach 2 is even more easy). Building a new release would be no problem.
> Some of my visions are:
> - clicking with the mouse to create new topics
> - creating of associations without the need of the drop down menu (like: http://morrisonpitt.com/jsPlumb/html/jquery/demo.html)
> - creating topics by just typing on the canvas
> - navigation in the canvas by using arrow keys
There are good ideas here.
How could they conceptualized further?
(How are associations created interactively in jsPlumb? I can't figure it out.)
> Lately I used FreeMind a lot again because it is so much
> faster to develop thoughts. Using the mindmap 10 minutes
> the same process would would take at least an hour in
> deepamehta. But it is missing the free positioning.
> So I would also like to look a new topic type that
> allows to create related or "sub" topics just by typing
> on the active topic in the canvas.
> And also it should be possible to show properties
> that are normally only visible in the detail pane
> in a mind map style fashion around the topic in the canvas.
> ... I am still thinking about this.
What are the interaction concepts that make FreeMind this effective and how could DM adapt to them?
Cheers,
Jörg
> Am 06.11.2010 15:38, schrieb Jörg Richter:
>> From my point of view the "grid positioning" facility would not be a perfect fit for the problem you describe. Grid positioning is meant to place an *arbitrary* number of topics. It always starts positioning *below* the topicmap's bottom-most topic and sets the viewport accordingly (by automatic scrolling). So, when you're working in an upper region of a large map and create a topic your viewport would be lost.
>>
>> To solve the overlapping problem 2 solutions come to my mind:
>>
>> 1) An algorithm that actually looks for a free position for placing a single topic *within* the current viewport (and would resort to scrolling only if the current viewport is filled-up).
>>
>> 2) An alternate gesture for creating a topic: right-click on the canvas and choose "Create Note" (or whatever type is selected in the upper toolbar) from the canvas's context-menu (similar to DM2). Thus, the user would have full control about the initial topic position *and* must not fumble with submenus (like in DM2).
>> The upper toolbar's "Create" button could remain as it is. So, there would be 2 alternate ways to do the same thing (the Create button and the context-menu), but with a reason: the Create button would still have the advantage of *being visible* as a lot of users would not intuitively right-click an empty spot (user feedback showed that).
>>
>> What do you think?
>> I could realize the changes quickly and provide a new release next week.
>>
>> Cheers,
>> Jörg
> Both approaches would directly address the problem you described in the first place (better placement for new topics). Both are sufficiently clear and could be implemented quite easily (approach 2 is even more easy). Building a new release would be no problem.
>
You are right, both of them are a good way to proceed with.
>> Some of my visions are:
>> - clicking with the mouse to create new topics
>> - creating of associations without the need of the drop down menu (like: http://morrisonpitt.com/jsPlumb/html/jquery/demo.html)
>> - creating topics by just typing on the canvas
>> - navigation in the canvas by using arrow keys
> There are good ideas here.
> How could they conceptualized further?
> (How are associations created interactively in jsPlumb? I can't figure it out.)
see this demo
http://morrisonpitt.com/jsPlumb/html/jquery/draggableConnectorsDemo.html
>> Lately I used FreeMind a lot again because it is so much
>> faster to develop thoughts. Using the mindmap 10 minutes
>> the same process would would take at least an hour in
>> deepamehta. But it is missing the free positioning.
>> So I would also like to look a new topic type that
>> allows to create related or "sub" topics just by typing
>> on the active topic in the canvas.
>> And also it should be possible to show properties
>> that are normally only visible in the detail pane
>> in a mind map style fashion around the topic in the canvas.
>> ... I am still thinking about this.
> What are the interaction concepts that make FreeMind this effective and how could DM adapt to them?
a) It can be navigated (navigation and content creation) purely by keyboard
b) a lot of information is visible right on the canvas without the need
to "open"
the topic and have a look at the detail pane (the new freemind
version does
have a detail pane in a separate html capable text window)
Yours,
Torsten
--
Torsten Ziegler
tor...@ziegi.de
> Oh, I do not think that it undermines the DM user interface paradigm,
> what I tried to say was that there might be some things to think of
> before starting approach number 2.
In case you like to tweak the canvas look & feel on your own: the current development version (github repos) provides plugin developers a mechanism to hook-in a custom topicmap rendering component. This component would completely replace the existing (HTML5 Canvas-based) component and would provide its own rendering (e.g. SVG-based) and event handling capabilities.
Have a look at topicmap_renderer.js (in the deepamehta3-client plugin)
The base class for all topicmap renderers is defined there. Custom implementation would override some or all of the adapter methods.
For the moment there are 2 topicmap renderers build on it:
- The standard HTML5 Canvas-based renderer (see topicmap-renderers/canvas.js)
- An OpenLayers-based geomap renderer (see the dm3-freifunk-geomap plugin)
A third topicmap renderer could be based on jsPlumb.
Plugins hook-in a custom topicmap renderer by the get_canvas_renderer() hook.
(see dm3_freifunk_geomap.js)
> For example it might be good to have a general infrastructure
> for acces to mouse and keyboard events and distributing them to the
> canvas or topics or other elemnts in deepamehta that could use them.
For the moment there is the topicmap renderer replacement mechanism described above.
Your custom renderer would be responsible for all rendering and event handling itself.
Sooner or later more common behavoir could migrate to the TopicmapRenderer base class (or hierarchy of base classes).
> You are right, both of them are a good way to proceed with.
I would do approach 2 in the next release then. The infrastructure is already there and it is quite an easy fix. (later on I would realize approach 1 as well).
I filed an issue:
https://github.com/jri/deepamehta3/issues#issue/15
Oh, pulling new associations out of topics via a dedicated handle!
That's nice!
We should do this in DM.
Question comes up: how to visualize that handle?
The current DM topic rendering mode (icon + label) is not really suited for that I suppose. However, DM's topic rendering demands improvement anyway (as you state below in b) and this seems to be an opportunity to kill 2 birds with one stone :-)
>> What are the interaction concepts that make FreeMind this effective and how could DM adapt to them?
> a) It can be navigated (navigation and content creation) purely by keyboard
Yes, we should have keyboard control.
I filed an issue:
https://github.com/jri/deepamehta3/issues#issue/14
The actual navigation approach is to be discussed.
(in a non-hierarchical graph there is no obvious approach I suppose)
> b) a lot of information is visible right on the canvas without the need to "open"
> the topic and have a look at the detail pane (the new freemind version does
> have a detail pane in a separate html capable text window)
Good point! We should do this.
Let's discuss this in a separate thread.
Again there is no obvious approach I suppose.
(Every topic type has different data fields. Field content can be very large.)
Thanks Torsten for your input!
Give me some time to realize the issues that are clear already.
I will start doing them in a week. Currently I work on Freifunk Geomap and Access Control.
Let me know if you need support with a jsPlumb topicmap renderer.
Cheers,
Jörg
without urging them to install or register which they don't want
Furthermore, the place where the imported icons are inserted, must be
more sophisticated, because currently, they keep popping up in the
center, cluttering your carefully positioned previous ones.
On Mon, 2010-11-01 at 04:19 -0700, x28 wrote:
Furthermore, the place where the imported icons are inserted, must be
more sophisticated, because currently, they keep popping up in the
center, cluttering your carefully positioned previous ones.
Works for me by now.