[ANN] Tycho - A PIM under development

ruby-talk
1

Hello, all.

I hate announcing things that don’t really exist yet, but I’m
wanting assistance on this. So here goes.

The idea for Tycho stems from my intense desire to have a good
Linux-based PIM (Personal Information Manager).

See http://tycho.rubyforge.org

Unlike many PIMs, this one is built around the concept of
random data – structureless, unformatted, context-limited –
the kind of data we get in everyday life. This will be the place
to store not just your phone numbers and addresses, but also
your ideas, quotations, account numbers, and anything else that
might go on a sticky note, a napkin, or an envelope.

Tycho is loosely based on an older version of Info Select, a
commercial product for Windows. (In its earliest incarnation,
it was a DOS program called Tornado.)

One of the things that made Tornado (or Info Select) special
was its incremental search capability. Even if you had a pile
of 1,000 unrelated notes, you could find the exact one you
were looking for – usually within five seconds.

I want to implement all of the most basic features and add some
metadata support. See the web page for details.

In my opinion, IS nowadays suffers from “feature bloat” – so it
is unlikely I would add all of its features even if I could.
What goes in and what stays out is, of course, partly a function
of who else gets interested in this project and wants to contribute
to it.

So there you have it. Please give me your comments, your suggestions,
and your code. :slight_smile:

Cheers,
Hal Fulton

2

Hal Fulton wrote:

Tycho is loosely based on an older version of Info Select, a
commercial product for Windows. (In its earliest incarnation,
it was a DOS program called Tornado.)

Wow - I used to run Info Select, I ran across an old copy the other day.
I immediately thought, “Hey, this is a lot like a Wiki.” Just wondered
if you’d thought about that, and how you think Tycho would differ from a
Wiki.

···

–
Chris
http://clabs.org/blogki

3

“Hal Fulton” hal9000@hypermetrics.com wrote in message
news:4037EDA2.3080501@hypermetrics.com…

Hello, all.

I hate announcing things that don’t really exist yet, but I’m
wanting assistance on this. So here goes.

The idea for Tycho stems from my intense desire to have a good
Linux-based PIM (Personal Information Manager).

See http://tycho.rubyforge.org

Wow! … would love to help but I don’t think I’ll be able to.

So naturally I am just brimming with suggestions :slight_smile:

Why limit the internal topic hierarchy to a tree? The tree limitation is
aggravating in PIMs, IE Bookmarks, other hierarchical “Favorites”, directory
structures, and the “tyrranny of the dominant decomposition” cited by
Aspects and Mdsoc.

Instead, merge topics and keywords/key-phrases. Allow each note to be tagged
with 1 or more topics (keywords/key-phrases), and allow those topics to form
a DAG of topic inclusion. So my note about the blues concert next month can
be tagged with “Blues” (which might be included in the topic “Music”), and
“March” (which might be included in the topic “Calendar”). Topic inclusion
is user-defined (or you can get fancy and guess at some, like <any valid

in “Calendar”). That way you can quickly categorize notes simply by
typing in some topics/keywords. Search still works fine, it’s easy to create
a tree view on the DAG, some manipulation operations may need some tweaking.
Perhaps the single tree in the tree view is just a projection of this DAG.

Mouse-less usability would be nice.

What GUI will you use?

…

4

Hal Fulton wrote:

Hello, all.

I hate announcing things that don’t really exist yet, but I’m
wanting assistance on this. So here goes.

I will like to assist.

The idea for Tycho stems from my intense desire to have a good
Linux-based PIM (Personal Information Manager).

I hear you! After having tried various PIM’s, Mindmappers, Outliners,
wikis (everyone has to write one at some time in his life) and more
exotic programs (like VKB from http://www.csdl.tamu.edu/VKB/) I feel
lost in a sea of unstructured yet structured thought.

Unlike many PIMs, this one is built around the concept of
random data – structureless, unformatted, context-limited –
the kind of data we get in everyday life. This will be the place
to store not just your phone numbers and addresses, but also
your ideas, quotations, account numbers, and anything else that
might go on a sticky note, a napkin, or an envelope.

Actually it somehow is all connected somehow even if the overall
structure is not easily recognized at first sight.

In my opinion, IS nowadays suffers from “feature bloat” – so it
is unlikely I would add all of its features even if I could.
What goes in and what stays out is, of course, partly a function
of who else gets interested in this project and wants to contribute
to it.

I think this calls for a ‘pluggable’ architecture. This is what makes
JEdit so attractive. The core team concentrates only on making the
Basics just right and let users add their favorite ‘pet’ function
themselves by giving them a well defined API.

By looking at the first picture on your website my hair stood on end,
because I have drawn the exact same picture just some weeks ago, when
you started this thread on rub-talk. I like tree-based information but I
also like sticky-notes. I tried using these desktop-sticky-notes that
float around always hidden under other windows, when you need them, and
there’s no way of grouping them or organising them. The cool thing about
sticky-notes is, that you can drag them around and visually group them,
which you could never do with any other PIM out there (not even with the
famous MindMappers). When using Unix I easily could create many virtual
desktops and fill them just with sticky-notes, but then theres no search
capability, well that’s what Tycho has to offer I guess.

I would be glad to help out if only to not do it myself (I think the
features that you suggest ‘click’ close enough with my ideas, to make
this happen).

One suggestion: how about moving this discussion to the rubyforge forum
(or create a tycho mailing list)?

Cheers,
Carsten.

5

I like the user interface, looks like it would be really nice to use.

Features I would like to see:

I would like to add fuzzy searches tothe list as I and I’m sure others
can’t spell. Also search for words near each other and search by word
list - as you start typing a word the words in the index that have those
letters appear in a list - for example if you type a, all the indexed
words starting with a are in the list. As you add letters the list
becomes smaller leading to the word you are looking for similar to the
index search
option in Windows.

Search results ranked by different options - most hits, hits in certain
fields, most matching words, newest documents, oldest documents and
probably others I haven’t thought of.

Import feature that knows different formats. For example I would like
to index my email - it should know that there are To:, From:, Subject:
lines etc in the mail and store the messages so they can be searched
by the automatically created indexes from the email template. Same
with pdf files, word docs etc. Also for the import option a date
qualifier
so for example you can do a import since a date for example when adding
your email to the database - you would like to add only new messages.
The same for indexing a web page or a set of notes. Or use the stored
time to see if the document has been updated since the last import.
If you have edited the database version and it has changed from where
you are importing from you might need to resolve how you are going to
handle the dual update - replace, add 2nd copy, merge or something like
a source code system like cvs where you have multiple changes to merge.

For an example of these type things you might look at the AskSam free
form database that does sort of what you are doing but without the
nice GUI interface ( http://www.asksam.com ). It only runs on Windows
so I’m looking for a linux alternative. Another set
of programs to look at for ideas might be agrep (approximate grep) and
the assocated indexing search program glimpse which also can be used for
searching
web pages ( http://webglimpse.net ). Unfortunately they now charge for
glimpse ( though for personal use it is cheap/free ), but even for
commercial use it is low priced. With glimpse you can search for words
and specify how many letters can be wrong (mis-spelled), boolean
searches,
word nearness etc.

Rick

···

–
Have you noticed that a “slight tax increase” costs us at least two
hundred dollars a year, but a “substantial tax cut” saves us maybe
thirty cents?

rickc@dallas.oilfield.slb.com

6

Chris Morris wrote:

Hal Fulton wrote:

Tycho is loosely based on an older version of Info Select, a
commercial product for Windows. (In its earliest incarnation,
it was a DOS program called Tornado.)

Wow - I used to run Info Select, I ran across an old copy the other day.
I immediately thought, “Hey, this is a lot like a Wiki.” Just wondered
if you’d thought about that, and how you think Tycho would differ from a
Wiki.

I think there are some similarities in concept.

Some differences in Tycho would be:

  1. not web-based
  2. not publicly accessible
  3. all local data
  4. much better interface and faster response - see #1 :slight_smile:

Would you be interested in helping with this?

Hal

7

Its Me wrote:

“Hal Fulton” hal9000@hypermetrics.com wrote in message
news:4037EDA2.3080501@hypermetrics.com…

Hello, all.

I hate announcing things that don’t really exist yet, but I’m
wanting assistance on this. So here goes.

The idea for Tycho stems from my intense desire to have a good
Linux-based PIM (Personal Information Manager).

See http://tycho.rubyforge.org

Wow! … would love to help but I don’t think I’ll be able to.

So naturally I am just brimming with suggestions :slight_smile:

Thanks for the positive feedback. Is it a lack of time that would
prevent you from contributing?

Why limit the internal topic hierarchy to a tree? The tree limitation is
aggravating in PIMs, IE Bookmarks, other hierarchical “Favorites”, directory
structures, and the “tyrranny of the dominant decomposition” cited by
Aspects and Mdsoc.

[snip]

I like the way you think. Perhaps later we can go in this direction,
but we have to start out relatively simply.

I do think that there could be different tree views, with this one
as the default.

If this were truly the 21st century (where are the flying cars?), I
would want to see an arbitrary graph produced in realtime with
hyperbolic viewing.

For that matter, I’ve always wanted to design a menu system where the
items were hexagons and the whole honeycomb centered on the mouse
pointer’s current location, with the most common selections nearer
the center. But that’s just me.

Mouse-less usability would be nice.

Definitely. I’ll add that as a constraint.

What GUI will you use?

For now, FXRuby. Later perhaps wxRuby or something.

Hal

8

Carsten Eckelmann wrote:

I will like to assist.

Glad to hear it, welcome! :slight_smile:

I think this calls for a ‘pluggable’ architecture. This is what makes
JEdit so attractive. The core team concentrates only on making the
Basics just right and let users add their favorite ‘pet’ function
themselves by giving them a well defined API.

That is a good idea. I would like to see all the most basic features
added before coding plugin capability, I think.

When we do make it plugin-capable, FreeRIDE’s infrastructure (called
FreeBASE – now separable from FreeRIDE) might be a good one to use.

By looking at the first picture on your website my hair stood on end,
because I have drawn the exact same picture just some weeks ago, when
you started this thread on rub-talk. I like tree-based information but I
also like sticky-notes. I tried using these desktop-sticky-notes that
float around always hidden under other windows, when you need them, and
there’s no way of grouping them or organising them. The cool thing about
sticky-notes is, that you can drag them around and visually group them,
which you could never do with any other PIM out there (not even with the
famous MindMappers). When using Unix I easily could create many virtual
desktops and fill them just with sticky-notes, but then theres no search
capability, well that’s what Tycho has to offer I guess.

Hmm, I had not thought of keeping track of notes’ positions. I have been
thinking in terms of IS, which basically puts them wherever it wants on
the screen. I like your idea, but how do we deal with the fact that only
a few notes on the stack will be visible at one time?

I would be glad to help out if only to not do it myself (I think the
features that you suggest ‘click’ close enough with my ideas, to make
this happen).

I’m sure we are not 100% in agreement, but we are close enough. And we
can compromise as needed. (I already compromised a little for Gavin
Sinclair, who doesn’t even have time to contribute, but would like to
be a user.)

One suggestion: how about moving this discussion to the rubyforge forum
(or create a tycho mailing list)?

I just created the tycho-devel mailing list, which should go into effect
in a few hours.

Please excuse me, but I am still learning how to admin a rubyforge
project.

Cheers,
Hal

9

I use Info Select, but would love to see a replacement too.
One I came across recently is FreeMind (http://freemind.sourceforge.net/).
If it were given the incremental search capability, it might be a nice replacement
to Info Select.
– Glenn

Carsten Eckelmann wrote:

···

Hal Fulton wrote:

Hello, all.

I hate announcing things that don’t really exist yet, but I’m
wanting assistance on this. So here goes.

I will like to assist.

The idea for Tycho stems from my intense desire to have a good
Linux-based PIM (Personal Information Manager).

I hear you! After having tried various PIM’s, Mindmappers, Outliners,
wikis (everyone has to write one at some time in his life) and more
exotic programs (like VKB from http://www.csdl.tamu.edu/VKB/) I feel
lost in a sea of unstructured yet structured thought.

Unlike many PIMs, this one is built around the concept of
random data – structureless, unformatted, context-limited –
the kind of data we get in everyday life. This will be the place
to store not just your phone numbers and addresses, but also
your ideas, quotations, account numbers, and anything else that
might go on a sticky note, a napkin, or an envelope.

Actually it somehow is all connected somehow even if the overall
structure is not easily recognized at first sight.

In my opinion, IS nowadays suffers from “feature bloat” – so it
is unlikely I would add all of its features even if I could.
What goes in and what stays out is, of course, partly a function
of who else gets interested in this project and wants to contribute
to it.

I think this calls for a ‘pluggable’ architecture. This is what makes
JEdit so attractive. The core team concentrates only on making the
Basics just right and let users add their favorite ‘pet’ function
themselves by giving them a well defined API.

By looking at the first picture on your website my hair stood on end,
because I have drawn the exact same picture just some weeks ago, when
you started this thread on rub-talk. I like tree-based information but I
also like sticky-notes. I tried using these desktop-sticky-notes that
float around always hidden under other windows, when you need them, and
there’s no way of grouping them or organising them. The cool thing about
sticky-notes is, that you can drag them around and visually group them,
which you could never do with any other PIM out there (not even with the
famous MindMappers). When using Unix I easily could create many virtual
desktops and fill them just with sticky-notes, but then theres no search
capability, well that’s what Tycho has to offer I guess.

I would be glad to help out if only to not do it myself (I think the
features that you suggest ‘click’ close enough with my ideas, to make
this happen).

One suggestion: how about moving this discussion to the rubyforge forum
(or create a tycho mailing list)?

Cheers,
Carsten.

10

Quoteing hal9000@hypermetrics.com, on Mon, Feb 23, 2004 at 01:47:35PM +0900:

>>Tycho is loosely based on an older version of Info Select, a
>>commercial product for Windows. (In its earliest incarnation,
>>it was a DOS program called Tornado.)
>
>Wow - I used to run Info Select, I ran across an old copy the other day.
>I immediately thought, "Hey, this is a lot like a Wiki." Just wondered
>if you'd thought about that, and how you think Tycho would differ from a
>Wiki.

I think there are some similarities in concept.

Some differences in Tycho would be:
1. not web-based
2. not publicly accessible
3. all local data
4. much better interface and faster response - see #1 :slight_smile:

I, too, was thinking how similar to a wiki this sounds, a few months ago
when this was first discussed.

The nice thing about designing this as a local front-end to a wiki is
that you can reuse a wiki's markup language, and its data storage, you
might even be able to use it to locally edit a live wiki.

And no, I have no time to help on this, I'm sorry. Also, I found
something like this for OS X, so don't need to build it:

  Untitled

  VoodooPad is a new kind of notepad. It's like having your own
  personal hypertext library, where you can jot down notes, web
  addresses, to-do lists... Anything on your mind. VoodooPad
  automatically links each page together, to form a miniature world wide
  web, on your desktop! Anybody familiar with the WikiWikiWeb will feel
  right at home with VoodooPad.

  Type in your notes, and highlight important words or phrases to create
  new pages; or drag and drop folders, images, applications, or URLs
  into VoodooPad - they're linked up whenever the word representing it
  is found.

I've started using it off-and-on, to see if I actually like it, and I need
some way to keep all my notes, so I thought I'd give it a try, I'm still
not sure its right for me.

The one thing I'd really, really, like that voodo pad doesn't do is
someway to graphically see the web of connections. I'd like to have
a pop-up window that showed each note as an icon (with title?), and
each link as a line, and as you select a note, it would show that note
in the view window.

I organize things in my head that way, and I've often wondered whether
or not I could organize things on a computer that way.

The problem with the web-like interfaces where you follow links is that
it can feel like you are in a maze - you see where you are, you see
where you can go, and you remember the last few places you were. What if
you could move up a level, see the relationships between places?

Cheers,
Sam

11

Hal Fulton wrote:

Carsten Eckelmann wrote:

I will like to assist.

Glad to hear it, welcome! :slight_smile:

I think this calls for a ‘pluggable’ architecture. This is what makes
JEdit so attractive. The core team concentrates only on making the
Basics just right and let users add their favorite ‘pet’ function
themselves by giving them a well defined API.

That is a good idea. I would like to see all the most basic features
added before coding plugin capability, I think.

When we do make it plugin-capable, FreeRIDE’s infrastructure (called
FreeBASE – now separable from FreeRIDE) might be a good one to use.

I just wanted to add my two cents in here.

If you are willing to invest in the learning curve of using FreeBASE, it
would be an excellent choice. But there would be a learning curve.

FreeBASE was developed as part of FreeRIDE which is, itself, entirely plugin
based. It was designed to separate and independent of FreeRIDE. Rich Kilmer
is using it in a couple other projects, and I also using it in another
project.

When we have time, we will break out FreeBASE into its own RubyForge
project. At the moment, you can read about it here:

http://freeride.rubyforge.org/wiki/wiki.pl?FreeBASE

Curt

12

Sam Roberts wrote:

The problem with the web-like interfaces where you follow links is that
it can feel like you are in a maze - you see where you are, you see
where you can go, and you remember the last few places you were. What if
you could move up a level, see the relationships between places?

c2.com has a bit of this built-in. For example, go here:
http://c2.com/cgi/wiki?ExtremeProgrammingSystem … then click the
VisualTour link at the bottom of the page.

···

–
Chris
http://clabs.org

13

Hal Fulton wrote:

Some differences in Tycho would be:

  1. not web-based
  2. not publicly accessible
  3. all local data
  4. much better interface and faster response - see #1 :slight_smile:

I have a private, local Wiki on my home machine. I’d guess this boils
down to interface then.

Would you be interested in helping with this?

Unfortunately, no. I think I’ll stick to tweaking out my own Wiki.

···

–
Chris
http://clabs.org

14

“Hal Fulton” hal9000@hypermetrics.com wrote in message
news:40399F36.4020501@hypermetrics.com…

So naturally I am just brimming with suggestions :slight_smile:

Thanks for the positive feedback. Is it a lack of time that would
prevent you from contributing?

Not just time … but that may change in the future.

I like the way you think. Perhaps later we can go in this direction,
but we have to start out relatively simply.

Point taken, tho’ I am not sure it will retrofit easily. One reference point
is Leo, a superb open-source Python project that is, in essense, a very easy
outline editor with a DAG underneath it’s outline view, and with an
additional layer of ‘literate programming’ functionality through functions
that traverse the structure and ‘weave’ a composite output from marked
sub-sections. For many (including me) it makes quite a passable PIM. Imho,
Leo is a great tool to look at for some useful ideas for Tycho
development.

And I heartily agree that a plug-in style is the best bet against bloat.

15

Glenn M. Lewis wrote:

I use Info Select, but would love to see a replacement too.
One I came across recently is FreeMind (http://freemind.sourceforge.net/).
If it were given the incremental search capability, it might be a nice
replacement
to Info Select.

Glenn,

I’ve looked at freemind, but not in great detail. My first impression
is that it is a little complex in terms of UI (and I’ve heard it’s
slow, though I don’t know).

I’ll probably go ahead with Tycho anyway, stealing ideas from other
places as appropriate.

Are you interested in contributing? :wink:

Hal

16

Sam Roberts wrote:

The problem with the web-like interfaces where you follow links is that
it can feel like you are in a maze - you see where you are, you see
where you can go, and you remember the last few places you were. What if
you could move up a level, see the relationships between places?

c2.com has a bit of this built-in. For example, go here:
http://c2.com/cgi/wiki?ExtremeProgrammingSystem … then click the
VisualTour link at the bottom of the page.

I cranked the width and depth, and couldn’t get more than a dozen nodes
on a page, it is very much still in the prototype stage! And it suffers
badly from the hideousness of html browsers as the lowest-common-
denominator gui.

The discussion links lead to something more promising:

It seems to ditch the web interface, but it sounds like they are having
problems using against a live wiki, the required xml-rpc interfaces
aren’t there yet. I’ll check back in a year, though, it looks like there
is a lot of interest.

Thanks for the pointer,
Sam

17

Hal Fulton wrote:

Glenn M. Lewis wrote:

I use Info Select, but would love to see a replacement too.
One I came across recently is FreeMind
(http://freemind.sourceforge.net/).
If it were given the incremental search capability, it might be a
nice replacement
to Info Select.

Glenn,

I’ve looked at freemind, but not in great detail. My first impression
is that it is a little complex in terms of UI (and I’ve heard it’s
slow, though I don’t know).

I’m using freemind a lot and I can tell you that it is not slow at all.
In fact it has a very good keyboard interface, you hardly have to use
the mouse at all. What is mising most though are features like arranging
notes visually in other formats than trees and of course having notes
appear in different contexts (which leo users will know as cloning).

I’ll probably go ahead with Tycho anyway, stealing ideas from other
places as appropriate.

Stealing is a hard word for “finding inspiration” isn’t it?

Cheers,
Carsten.

PS: how about moving this conversation to the tycho mailing list?
http://rubyforge.org/mailman/listinfo/tycho-devel