Seek: faceted browsing in Thunderbird
David Huynh
I've made an extension that adds faceted browsing to Thunderbird:
Screencast: http://www.vimeo.com/758863/
Website: http://simile.mit.edu/seek/
I've already been in conversation with David Ascher and Wayne Merry, but
I'd still love to get some more feedback from people here!
Thanks!
David
Kent James
>
> I've made an extension that adds faceted browsing to Thunderbird
>
I've only reviewed your video, but it seems to me to be an outstanding
addition to TB. It also highlights an important issue for Thunderbird's
future that has concerned me.
There is an understandable desire to make Thunderbird a complete
personal messaging client, perhaps with the functionality of Outlook
circa 2000. But Outlook does not sit still, and there is no real hope of
every being a better Outlook than Outlook.
At the same time, there is a lot of research going on now in unified
information management, such as MIT projects like Haystack, or the SRI
CALO project (for example openiris). The Mozilla platform, with its open
nature, cross-platform capability, and extension support, could be the
platform of choice for implementation of advanced concepts of
information management. If Mozilla was routinely adopted by the leading
edge of information management research, then the possibilities of
leapfrogging Outlook would start to materialize.
I'n very encouraged that you have decided to implement one of the
innovations coming out of MIT in Thunderbird. I hope we can see more
collaboration in the future between Thunderbird and the research community.
Alan Lord
Before I try it though I have a question or two?
1, I have many separate email accounts in TB each with own inbox. Will
your seek tool give me the ability to search through all of them at the
same time? (This is one my biggest issues with the current TB search
facility. I have to manually trawl through every mailbox).
2. Are the individual facets user configurable?
Thanks, the tools like a great idea - all I need now is a bigger screen ;-)
Alan
David Huynh
Kent,
Indeed, that is precisely the reason why I made Seek on Thunderbird. I
have been, and still am, very amazed at how well Firefox fits as a
research platform (seriously, I could not have finished my PhD working
on IE--ew...). So I think there is a huge potential for Thunderbird to
do the same.
Seek is the first extension I ever wrote for Thunderbird (I've written 3
for Firefox). And I wanted to go all out--beyond simple incremental
improvements to it--so that I can point my colleagues at MIT to
Thunderbird and tell them it's worth investing efforts on.
Regarding Outlook, I dumped Outlook 6 years ago and never looked back.
If there is no integration with Exchange, there's really no value that
Outlook offers me above Thunderbird. Thunderbird just "feels" so much
better to use. And as you say, if you can garner enough interests from
the research community, Thunderbird leapfrogging Outlook in terms of
innovations is not a doubt in my mind.
By the way, Seek took me (one person) 5 weeks part-time to implement. I
think you can use it (and the screencast) to make your point about
Thunderbird's future.
This is very exciting!
David
David Huynh
> David Huynh wrote:
>> Hi all,
>>
>> I've made an extension that adds faceted browsing to Thunderbird:
>>
>> Screencast: http://www.vimeo.com/758863/
>> Website: http://simile.mit.edu/seek/
>>
>> I've already been in conversation with David Ascher and Wayne Merry,
>> but I'd still love to get some more feedback from people here!
>>
>> Thanks!
>>
>> David
> That looks really cool David!
>
> Before I try it though I have a question or two?
>
> 1, I have many separate email accounts in TB each with own inbox. Will
> your seek tool give me the ability to search through all of them at the
> same time? (This is one my biggest issues with the current TB search
> facility. I have to manually trawl through every mailbox).
Not at the moment. I'm investigating how to make it work over a virtual
folder, which can combine several other folders (see my message in this
forum titled "enumerating messages for a virtual folder"). I need some
technical help with that--if you know someone who can answer my
question, that would be awesome.
> 2. Are the individual facets user configurable?
How do you want to configure them?
> Thanks, the tools like a great idea - all I need now is a bigger screen ;-)
Hehe :) Yeah, I still need to optimize the screen space usage.
alta88
David,
On a fast try, all I can say is wow. It's rare to see this level of UI
design in a mozilla extension, very nice. Seek has initial trouble on
account level folders, and doesn't seem to recover, nor did I notice a
way to do 'root' ie all accounts. Not sure how/if it's extensible to
subfolders.
But on a 10700 message newsgroup (Tb support) it indexed in about a
minute and performed pretty well. Next to try is a very large RSS
folder - these non pop/imap folders are a good way to test flexibility,
which will be required if Moz Messaging carries through on the vision of
different message types in Tb..
Glad you chose to prove the concept in Tb. All you need is a button now ;)
Philip Reames
> I've made an extension that adds faceted browsing to Thunderbird:
> I've already been in conversation with David Ascher and Wayne Merry, but
> I'd still love to get some more feedback from people here!
I'm going to be using this extension all the time. Thank you!
Two suggestions:
- Recursive indexing - If I'm search my inbox, I usually want everything
in the folders under it as well. Except when I don't, of course. :)
- Local Folders - Once recursive indexing is an option, it would be
really nice to be able to search from the root of Local Folders. I'm
guessing this was what you're earlier post was about?
Once again, wow! I love this extension.
Yours,
Philip Reames
Cory Albrecht
That's got to be one of the cooler extension I've seen for TB! I never
even thought of applying that type of search to email. :-) TB's pitiful
filtering capabilities it something I've been disappointed in, and I
can't wait until my personal list of make-or-break extensions will run
in TB3 which seems to have better filters.
I just have two other comments on Seek.
1) With Seek running, I could not delete any emails. I kept getting a
pop-up saying "Not implemented yet". I had to disable Seek and restart
TB to be able to delete.
2) Is there a way to change the column size in the different panels
other than the automatic change that happens when you change the width
if the panel itself? I couldn't click and drag on the dividers in the
column headers to resize them.
David Huynh
>
>
>
> ---On 2008.Mar.08 10:20 AM, David Huynh wrote:
>> Hi all,
>>
>> I've made an extension that adds faceted browsing to Thunderbird:
>>
>> Screencast: http://www.vimeo.com/758863/
>> Website: http://simile.mit.edu/seek/
>>
>> I've already been in conversation with David Ascher and Wayne Merry,
>> but I'd still love to get some more feedback from people here!
>>
>> Thanks!
>>
>> David
>
> David,
>
> On a fast try, all I can say is wow. It's rare to see this level of UI
> design in a mozilla extension, very nice. Seek has initial trouble on
> account level folders, and doesn't seem to recover, nor did I notice a
> way to do 'root' ie all accounts. Not sure how/if it's extensible to
> subfolders.
Ah, I haven't really tested it on account level folders.
> But on a 10700 message newsgroup (Tb support) it indexed in about a
> minute and performed pretty well. Next to try is a very large RSS
> folder - these non pop/imap folders are a good way to test flexibility,
> which will be required if Moz Messaging carries through on the vision of
> different message types in Tb..
By the way, the index is just a Javascript data structure... So, it
actually won't scale that much. I will have to re-do it using sqlite.
Sqlite will also let me persist the indices so that Seek won't have to
reindex every time you go back to a folder.
I actually have another technical question: what's a surefire way of
registering for updates on folders so that I can make sure the indices
are in sync with Thunderbird's native stores of the messages? I want to
keep track of changes to a folder even if it's not selected in the
folder tree and displayed in the thread tree.
David Huynh
> David Huynh wrote:
>> I've made an extension that adds faceted browsing to Thunderbird:
> <snip>
>> I've already been in conversation with David Ascher and Wayne Merry,
>> but I'd still love to get some more feedback from people here!
> Wow. Within 30 seconds of playing around with this, I can already tell
> I'm going to be using this extension all the time. Thank you!
>
> Two suggestions:
> - Recursive indexing - If I'm search my inbox, I usually want everything
> in the folders under it as well. Except when I don't, of course. :)
I think that's quite do-able with some technical help from you
Thunderbird gurus... :)
> - Local Folders - Once recursive indexing is an option, it would be
> really nice to be able to search from the root of Local Folders. I'm
> guessing this was what you're earlier post was about?
Yup.
> Once again, wow! I love this extension.
Glad you like it :) It really shows the power of the platform--that's
about 4 ~ 5 weeks of work.
David
David Huynh
> David Huynh wrote, on 2008-03-08 12:20:
>> Hi all,
>>
>> I've made an extension that adds faceted browsing to Thunderbird:
>>
>> Screencast: http://www.vimeo.com/758863/
>> Website: http://simile.mit.edu/seek/
>>
>> I've already been in conversation with David Ascher and Wayne Merry,
>> but I'd still love to get some more feedback from people here!
>
> That's got to be one of the cooler extension I've seen for TB! I never
> even thought of applying that type of search to email. :-) TB's pitiful
> filtering capabilities it something I've been disappointed in, and I
> can't wait until my personal list of make-or-break extensions will run
> in TB3 which seems to have better filters.
>
> I just have two other comments on Seek.
>
> 1) With Seek running, I could not delete any emails. I kept getting a
> pop-up saying "Not implemented yet". I had to disable Seek and restart
> TB to be able to delete.
Actually if you "disengage" Seek (upper right button), or select another
folder, then you get back Thunderbird's controls. Alternatively, you can
drag and drop messages to Trash to delete.
You see, I was trying to copy C++ implementation over to Javascript and
I wasn't sure what would work and what wouldn't... Delete was a
dangerous thing so I opted for leaving it out for now.
BTW, the code is here:
http://simile.mit.edu/repository/seek/trunk/
> 2) Is there a way to change the column size in the different panels
> other than the automatic change that happens when you change the width
> if the panel itself? I couldn't click and drag on the dividers in the
> column headers to resize them.
Oh, ... I didn't realize that. I'll see if I can fix it.
David
alta88
---On 2008.Mar.08 04:25 PM, David Huynh wrote:
>
>
> I actually have another technical question: what's a surefire way of
> registering for updates on folders so that I can make sure the indices
> are in sync with Thunderbird's native stores of the messages? I want
> to keep track of changes to a folder even if it's not selected in the
> folder tree and displayed in the thread tree.
>
>
what was impressive about the indexing was that you handled an edge case
really elegantly with the search time countdown. so it didn't even feel
like a minute..
i believe you would need to add a folder listener to a session. do a
search on nsIMsgMailSession or nsIFolderListener or AddFolderListener.
one additional (somewhat tangential) thing which would really launch
this is everyone's favorite request: regex search and complex searches.
David Huynh
>
>
>
> ---On 2008.Mar.08 04:25 PM, David Huynh wrote:
>>
>>
>> I actually have another technical question: what's a surefire way of
>> registering for updates on folders so that I can make sure the indices
>> are in sync with Thunderbird's native stores of the messages? I want
>> to keep track of changes to a folder even if it's not selected in the
>> folder tree and displayed in the thread tree.
>>
>>
>
> what was impressive about the indexing was that you handled an edge case
> really elegantly with the search time countdown. so it didn't even feel
> like a minute..
Hope you liked the live drag and drop support for rearranging the
facets, too. I don't remember seeing that kind of drag and drop live
feedback in Mozilla apps before.
> i believe you would need to add a folder listener to a session. do a
> search on nsIMsgMailSession or nsIFolderListener or AddFolderListener.
Ah, thanks for the pointer. I'll dig around.
> one additional (somewhat tangential) thing which would really launch
> this is everyone's favorite request: regex search and complex searches.
That would depend more on Thunderbird's native content text search...
David
bhar...@gmail.com
David,
Sweet extension. Definately very useful.
(I'll echo a gratuitous feature request ... the ability for seek to
remember the indices of multiple folders, and update as message are
received/deleted/sorted)
(..and #2... :-). A way to "minimize" seek --> ie. go from a large
rectangle that is the seek UI to a think strip and back).
Anyway, fantastic.
Cheers!
David Huynh
Yes, definitely. My next task is to investigate how to use sqlite to
keep the indices around and also to scale better.
One technical question: is there a way to tell when a folder was last
updated? The problem I'm expecting is that Seek gets disabled or
uninstalled and then when it gets enabled/installed again it keeps
thinking that its indices are up to date but they are not, because Seek
wasn't getting any event when it was not activated.
> (..and #2... :-). A way to "minimize" seek --> ie. go from a large
> rectangle that is the seek UI to a think strip and back).
Try Ctrl-K or Command-K.
I think if I can get Seek to work over several folders at once, or over
your whole email-base, then the UI optimized for its use will be
substantially different. The folder tree might even become a facet--each
folder has a checkbox so you can include/exclude it from the search.
The current UI paradigm for all email programs (except GMail) is very
much the same as the UI paradigm for the conventional file system. You
have a tree of folders and then selecting a folder gives you its content
items. Now if we're transiting over to think of email as being stored in
a database instead, where each message might be tagged with multiple
tags and hence there is no single strict hierarchy, then the UI paradigm
must change, quite radically...
-----
In my ideal email program, email messages get backed up automatically
into folders by date, perhaps monthly. This is just a performance
optimization, but it also facilitates back-up. But something like Seek
would let me search over the whole email-base at any time. Perhaps
there's a timeline facet where I can specify roughly how much time to go
back (a month, a year,...). The timeline is also automatically marked
with events, e.g., "start of grad school", "joe's wedding", "annie's
first baby", "trip to france", that are somehow learned from the content
of the email.
I can tag not just the email messages, but also the correspondents.
"Joe" is "friend", "from MIT", "play soccer", "entrepreneur",... And
then I can faceted browse based on the tags on the correspondents.
Some of those ideas have been explored in research before. But none has
made it into a program that I can confidently use and rely on daily. I
have wanted such a thing for 7 years now, and it's time I'm rolling up
my sleeves and do something about it. :)
David
David Ascher
>
> In my ideal email program, [...]
I love this part of the email.
People intrigued by this should also watch the Xobni demo on xobni.com.
Xobni is an Outlook plugin, but they tackle some of the same issues that
many have talked about as possible avenues for Thunderbird.
--david
ovidiu
> much the same as the UI paradigm for the conventional file system. You
> have a tree of folders and then selecting a folder gives you its
> content items. Now if we're transiting over to think of email as being
> stored in a database instead, where each message might be tagged with
> multiple tags and hence there is no single strict hierarchy, then the
> UI paradigm must change, quite radically...
less like gmail, should provide more friendly and natural controls to
the user, it seams to be that not so natural [but computerized] folder
tree approach that became a "standard" and most of the people have a
problem in getting out of it. Proof is that gmail imap presents folders
corresponding to it's labels in all apps. I prefer using tags and alike,
but hmm ..
So I just wonder what it takes to invite people out of that ..
>
> -----
>
> In my ideal email program, email messages get backed up automatically
> into folders by date, perhaps monthly. This is just a performance
> optimization, but it also facilitates back-up. But something like Seek
> would let me search over the whole email-base at any time. Perhaps
> there's a timeline facet where I can specify roughly how much time to
> go back (a month, a year,...). The timeline is also automatically
> marked with events, e.g., "start of grad school", "joe's wedding",
> "annie's first baby", "trip to france", that are somehow learned from
> the content of the email.
>
> I can tag not just the email messages, but also the correspondents.
> "Joe" is "friend", "from MIT", "play soccer", "entrepreneur",... And
> then I can faceted browse based on the tags on the correspondents.
>
> Some of those ideas have been explored in research before. But none
> has made it into a program that I can confidently use and rely on
> daily. I have wanted such a thing for 7 years now, and it's time I'm
> rolling up my sleeves and do something about it. :)
>
say that after such an example and it all starts to glow in a different
light
now that's really great
what can I say more ..
Caspy7
Seek is on/engaged and I double click on a message to open it up in a
new window. The window opens, but nothing is displayed.
ovidiu
If only I take it as a more natural UI for the advanced search (native
TB) and it still proves a step forward in the "improved search" goal
stated some time ago in a road map scheme here ..
the only thing is that, when giving away such great stuff, people will
have incredibly higher expectations from you now :) [Like playing with
calendar data also ? ha ..]
David Huynh
>> The current UI paradigm for all email programs (except GMail) is very
>> much the same as the UI paradigm for the conventional file system. You
>> have a tree of folders and then selecting a folder gives you its
>> content items. Now if we're transiting over to think of email as being
>> stored in a database instead, where each message might be tagged with
>> multiple tags and hence there is no single strict hierarchy, then the
>> UI paradigm must change, quite radically...
> This is a very delicate issue IMHO. As I think the db approach, more or
> less like gmail, should provide more friendly and natural controls to
> the user, it seams to be that not so natural [but computerized] folder
> tree approach that became a "standard" and most of the people have a
> problem in getting out of it. Proof is that gmail imap presents folders
> corresponding to it's labels in all apps. I prefer using tags and alike,
> but hmm ..
>
> So I just wonder what it takes to invite people out of that ..
Agreed.
I personally am not so concerned with other people at the moment--I just
want a solution that works for me. Then, using that solution, I can
start to generalize it and explore how to make it easier for other
people to adapt. Call it incubation :)
This is a big problem space, and I think if I start by trying to please
everyone, I'd just get scared to even start at all :)
David Huynh
I had seen Xobni before, but it didn't look like it would meet my own needs.
I wonder if it was a design decision for Xobni to stick everything in a
slim sidebar or it was impossible for them to integrate with the rest of
Outlook's UI.
David
David Huynh
Ah, thanks for spotting that bug!
(I replaced the default tree view of the thread tree with my own and
that's quite a beast to duplicate. It'd be nice if that thread tree is a
little easier to override partially.)
David
David Huynh
> David Huynh wrote:
>> Hi all,
>>
>> I've made an extension that adds faceted browsing to Thunderbird:
>>
>> Screencast: http://www.vimeo.com/758863/
>> Website: http://simile.mit.edu/seek/
>>
>> I've already been in conversation with David Ascher and Wayne Merry,
>> but I'd still love to get some more feedback from people here!
>>
>> Thanks!
>>
>> David
> great extension
> If only I take it as a more natural UI for the advanced search (native
> TB) and it still proves a step forward in the "improved search" goal
> stated some time ago in a road map scheme here ..
I think there is a difference between "searching" and "browsing". When
you are faced with a search text box or a search dialog box, you need to
know what text to enter. When you are faced with browsing options, you
only need to pick among the choices already there. The former is
"recall" and the latter is "recognition". Recognition is easier to do
mentally than recall. This is a well studied research topic.
Teevan, Alvarado, Ackerman, Karger. The Perfect Search Engine Is Not
Enough: A Study of Orienteering Behavior in Directed Search. CHI 2004.
http://people.csail.mit.edu/teevan/work/publications/papers/chi04.pdf
The camera shopping example I used in the screencast helps illustrate
this. Looking at
http://reviews.cnet.com/digital-cameras/?tag=promo.dc
it's quite easy to pick a manufacturer. If instead I had to type into a
text box the name of the manufacturer I want, I risk misspelling, and I
also miss out on names that I don't know or don't think they make
cameras. Even better, looking at that web page, I know immediately that
Canon has the most cameras reviewed without clicking on anything. I know
how many results a click will get me, saving me from heading down a dead
end. (iTunes doesn't have preview counts :( .)
Searching and browsing are distinct but complementary, and I believe
that we need good tools/UIs for both. I happen to have some expertise in
building browsing UIs that I hope to contribute here.
> the only thing is that, when giving away such great stuff, people will
> have incredibly higher expectations from you now :) [Like playing with
> calendar data also ? ha ..]
Better to face high expectations than obscurity... :)
David
Robert Kaiser
> The current UI paradigm for all email programs (except GMail) is very
> much the same as the UI paradigm for the conventional file system. You
> have a tree of folders and then selecting a folder gives you its content
> items. Now if we're transiting over to think of email as being stored in
> a database instead, where each message might be tagged with multiple
> tags and hence there is no single strict hierarchy, then the UI paradigm
> must change, quite radically...
Interestingly, that "folders vs. tag-mashup" thing is coming up again
and again for different topics and in different ways. Interestingly,
there are significant groups of people who only feel confident in being
able to work with one of them, for both approaches.
I more and more realize this is probably something that goes way back
into the personality of users, how they organize themselves in most of
their lives actually, and that killing off one or the other can never
lead to a really satisfied user base.
Any software going into organizing information probably needs to allow
for both hierarchical and flat-but-annotated structures for being
considered well-organized by all both of those groups of human
organizing themselves.
Robert Kaiser
bhar...@gmail.com
>
> - Show quoted text -
The way you are thinking about the email is interesting. What caught
my attention more is the Contact/Address-book slant. I've been trying
to figure out a way to tag/categorize my friends/work collegues/
relatives better than putting that information into a "Company" or
etc... tag in Thunderbird/Outlook. Being about to tag contacts (or
categorize or whatever you want to name the action) would be IMMENSELY
helpful.
As for the timeline/event based view. That's interesting. I think it
would more helpful is the two way street between timeline and
faceting...
ie. (a) i could facet my emails....then do view a timeline about the
faceted onces...this would be great way to visually pinpoint that
email about that meeting two weeks ago....
and (b) i could timeline my emails...then facet....we were discussing
something important via a month ago.....the facet by tag (work /
project x)...
Interestinig ideas....(and I can't wait for you to do it ...so I get
the side benefit of making my life easier ;-)).
cheers,
Bryan W Clark
> David Ascher wrote:
>> David Huynh wrote:
>>>
>>> In my ideal email program, [...]
>>
>> I love this part of the email.
>>
>> People intrigued by this should also watch the Xobni demo on
>> xobni.com. Xobni is an Outlook plugin, but they tackle some of the
>> same issues that many have talked about as possible avenues for
>> Thunderbird.
>>
>
> I had seen Xobni before, but it didn't look like it would meet my own
> needs.
I really like the Xobni demo, they do tackle some really interesting and
fun problems by looking at email from (what I would call) a more person
oriented approach.
> I wonder if it was a design decision for Xobni to stick everything in a
> slim sidebar or it was impossible for them to integrate with the rest of
> Outlook's UI.
Good question. I like how it gives you the ability to do searching /
browsing within the sidebar so you can keep your current email/task
focused and I was wondering the same thing about it.
You could still solve the mini researcher problem without being a
sidebar, but it's not a bad way to do it.
~ Bryan
Wayne Mery
> David Huynh wrote:
>> The current UI paradigm for all email programs (except GMail) is very
>> much the same as the UI paradigm for the conventional file system. You
>> have a tree of folders and then selecting a folder gives you its content
>> items. Now if we're transiting over to think of email as being stored in
>> a database instead, where each message might be tagged with multiple
>> tags and hence there is no single strict hierarchy, then the UI paradigm
>> must change, quite radically...
>
> Interestingly, that "folders vs. tag-mashup" thing is coming up again
> and again for different topics and in different ways. Interestingly,
> there are significant groups of people who only feel confident in being
> able to work with one of them, for both approaches.
I get that sense as well. And I am one of them. I had been
labeling/tagging my bugs and other types of mail with custom tags for a
couple years. But several aspects are of tagging are limiting:
1. one often wants to see mails that are outside the tag group (example
new message comes in that matches a tagged message - if your view is the
tag, and the new message isn't tagged then you don't see the new message)
2. changing custom searches feels tramatic/you lose your place
3. tags in saved searches is helpful, but virtual folders are not
robust, plus you can't thread
4. some other things I can't put my finger on at the moment
plus, folder names easy to remember, can be organized/hierarchical and
plastered everywhere in the UI.
So I have pretty much left tags and gone to:
- for big projects separate folders for archive/done, waiting response,
pending my action/unread, etc - within each project area
- virtual folders to aggregate across a project or account
- multiple UI -- tabbed message window when using trunk, multiple 3pane
when not using trunk, thunderbird on two PCs
I still don't consider this ideal. And I haven't ruled out the idea of
readopting tags.
So Robert makes a good point. And speaking more generally, perhaps when
thinking about the product (and new function in particular) there could
be more thinking outside the "technical" realm in which most of us are
most comfortable, and reach (and even test) more into the user realm and
give usage style/types/habits greater attention and consideration.
Indeed, more thinking outside the box could be helpful - sometimes
considering the "impossible" helps deliver something unexpected,
refreshing and powerful.
Wayne
ovidiu
> ovidiu wrote:
>> So I just wonder what it takes to invite people out of that ..
>
> Agreed.
>
> I personally am not so concerned with other people at the moment--I
> just want a solution that works for me.
to indicate that such approach may just do it. ;)
> Then, using that solution, I can start to generalize it and explore
> how to make it easier for other people to adapt. Call it incubation :)
>
> This is a big problem space, and I think if I start by trying to
> please everyone, I'd just get scared to even start at all :)
>
more natural than in some constrained environments with lots of
marketing and biz factors around .. or at least it looks like it should
be like that ..
ovidiu
[like browsing, using things "on the table"] is expected in other areas
that are not immediately related to searches and browsing. Like filters,
which are a great feature, but still a bit cryptic to some normal users
that don't have the time to play mind games with thorough logic
arrangements of them in a panel with some operators ..
So, what I'm saying is, maybe, "seek" could be an inspiration for the UI
approach for other features that may be great already, but not fully
understood and used at large cause are somehow "advanced" looking. To
follow the same example of filters, simply having to check some boxes in
couple of side by side little panels like in "seek" coud make a
difference in comparison to the dropdown today, at least for being more
self explanatory if not for the ease of general use.
Call it browsing for filter definitions or anything and there you have it ..
> I happen to have some expertise in building browsing UIs that I hope
> to contribute here.
I hope that too :)
>> the only thing is that, when giving away such great stuff, people
>> will have incredibly higher expectations from you now :) [Like
>> playing with calendar data also ? ha ..]
>
> Better to face high expectations than obscurity... :)
+1
ovidiu
perspective:
1. habit. Once you learned one way you just go back to doing the job,
while any "new and better way" have to be already far better to be worth
spending time on it
2. confidence (in the tool ans in yourself ). Once a process riches a
level of automation from the user perspective, another one may just be
undermining confidence in one's own skills along with having to deal
with building trust in a new tool. At least for taking a bit of time to
get things to the same speed. And nevertheless, is the "hands on"
phenomena . Meaning that you can have things more or less organized on
shelves and manage them, natural or not so (let's say the folder
structure), or can have somebody do it for you :) (searches etc). In
this case, TB. It seams easier, but involves a confidence level.
I observed this in various ways when helping people migrate soft and
even when teaching to finally use one thing as it was originally
designed to work (against some mis usage people had developed with no
support).
[ For another example, the "layers", present in all CAD and graphic
tools today, are hard to imagine "intersecting" with other not so
"realistic" attributes because of their strong equivalent in real world
(classic layers, onion skinning, tracing paper etc). While they are
really only attributes themselves. Or call them labels, tags .. But for
a user to digest that, you don't go to dissect the IT attributes or
chemistry behind, but is rather about a way of presenting it further in
correspondence with some "friendly" procedures. Well, some strong habits
or real world resemblance may give you a hard time there .. ]
An easy response is have a good self explanatory Ui. And keep it simple.
But is complicated to keep it simple ..
Andrew Sutherland
> I wonder if it was a design decision for Xobni to stick everything in a
> slim sidebar or it was impossible for them to integrate with the rest of
> Outlook's UI.
I suspect it's the impossible one. It's been a while, but at one point
in my pursuit of e-mail visualization I was trying to do so from within
Outlook. Outlook did not appear to provide a mechanism for UI
augmentation apart from adding toolbar buttons or entirely replacing a
folder's content area with your own. My plan for integration was going
to be to muck with the internal window hierarchy and hope Outlook didn't
stop me...
Andrew
David Huynh
> On 3/11/2008 12:29 PM, Robert Kaiser wrote:
>> David Huynh wrote:
>>> The current UI paradigm for all email programs (except GMail) is very
>>> much the same as the UI paradigm for the conventional file system. You
>>> have a tree of folders and then selecting a folder gives you its content
>>> items. Now if we're transiting over to think of email as being stored in
>>> a database instead, where each message might be tagged with multiple
>>> tags and hence there is no single strict hierarchy, then the UI paradigm
>>> must change, quite radically...
>>
>> Interestingly, that "folders vs. tag-mashup" thing is coming up again
>> and again for different topics and in different ways. Interestingly,
>> there are significant groups of people who only feel confident in being
>> able to work with one of them, for both approaches.
>
> I get that sense as well. And I am one of them. I had been
> labeling/tagging my bugs and other types of mail with custom tags for a
> couple years. But several aspects are of tagging are limiting:
>
> 1. one often wants to see mails that are outside the tag group (example
> new message comes in that matches a tagged message - if your view is the
> tag, and the new message isn't tagged then you don't see the new message)
Right. That's why there are the "include whole threads" and "include new
messages" in Seek. That's actually my #1 feature for myself.
> 2. changing custom searches feels tramatic/you lose your place
> 3. tags in saved searches is helpful, but virtual folders are not
> robust, plus you can't thread
> 4. some other things I can't put my finger on at the moment
> plus, folder names easy to remember, can be organized/hierarchical and
> plastered everywhere in the UI.
So, there are 2 issues
- the data model (associative tagging vs. folder-based strict
hierarchical containment)
- the user interface
You can in fact choose the tagging data model but expose a hierarchical
containment interface. The advantage is of course that you can then
break that strict hierarchical containment interface at any time and
morph toward a tagging interface.
There's actually nothing to prevent imposing a hierarchy of tags
themselves. And then the UI would look pretty much like the conventional
folder-based UI. And tags can be easy to remember, too.
> So I have pretty much left tags and gone to:
>
> - for big projects separate folders for archive/done, waiting response,
> pending my action/unread, etc - within each project area
> - virtual folders to aggregate across a project or account
> - multiple UI -- tabbed message window when using trunk, multiple 3pane
> when not using trunk, thunderbird on two PCs
>
> I still don't consider this ideal. And I haven't ruled out the idea of
> readopting tags.
>
> So Robert makes a good point. And speaking more generally, perhaps when
> thinking about the product (and new function in particular) there could
> be more thinking outside the "technical" realm in which most of us are
> most comfortable, and reach (and even test) more into the user realm and
> give usage style/types/habits greater attention and consideration.
> Indeed, more thinking outside the box could be helpful - sometimes
> considering the "impossible" helps deliver something unexpected,
> refreshing and powerful.
I don't know what Mozilla Messaging's own timeframe is, but I'd be
willing to explore some of these concepts in the form of extensions and
then let you folks be the judge of what's good and what's bad...
Sometimes, without actually holding the artifact in your hand, it's hard
to really imagine what it would be like.
David
David Huynh
> Robert Kaiser wrote:
>> [snip]
> IMHO this "folders vs. tag-mashup" is related to two factors from users
> perspective:
> 1. habit. Once you learned one way you just go back to doing the job,
> while any "new and better way" have to be already far better to be worth
> spending time on it
> 2. confidence (in the tool ans in yourself ). Once a process riches a
> level of automation from the user perspective, another one may just be
> undermining confidence in one's own skills along with having to deal
> with building trust in a new tool. At least for taking a bit of time to
> get things to the same speed. And nevertheless, is the "hands on"
> phenomena . Meaning that you can have things more or less organized on
> shelves and manage them, natural or not so (let's say the folder
> structure), or can have somebody do it for you :) (searches etc). In
> this case, TB. It seams easier, but involves a confidence level.
Totally agreed. We're fighting ingrained habits.
> I observed this in various ways when helping people migrate soft and
> even when teaching to finally use one thing as it was originally
> designed to work (against some mis usage people had developed with no
> support).
>
> [ For another example, the "layers", present in all CAD and graphic
> tools today, are hard to imagine "intersecting" with other not so
> "realistic" attributes because of their strong equivalent in real world
> (classic layers, onion skinning, tracing paper etc). While they are
> really only attributes themselves. Or call them labels, tags .. But for
> a user to digest that, you don't go to dissect the IT attributes or
> chemistry behind, but is rather about a way of presenting it further in
> correspondence with some "friendly" procedures. Well, some strong habits
> or real world resemblance may give you a hard time there .. ]
>
> An easy response is have a good self explanatory Ui. And keep it simple.
> But is complicated to keep it simple ..
:) Yeah. Simplicity is complicated. You can say that again.
David
David Huynh
>
> The way you are thinking about the email is interesting. What caught
> my attention more is the Contact/Address-book slant. I've been trying
> to figure out a way to tag/categorize my friends/work collegues/
> relatives better than putting that information into a "Company" or
> etc... tag in Thunderbird/Outlook. Being about to tag contacts (or
> categorize or whatever you want to name the action) would be IMMENSELY
> helpful.
The idea here is that you have substantially fewer correspondents than
email messages, and so tagging correspondents is easier than tagging
messages. You can extend this idea further. For example, instead of
tagging correspondents, tag their domains (e.g., "mit.edu" is
"education", "professional", "usa", "cambridge").
This is a departure from thinking of email stuff as a flat list or a
strict tree of the same kind of object, to thinking of email stuff as a
graph of related things of different kinds.
> As for the timeline/event based view. That's interesting. I think it
> would more helpful is the two way street between timeline and
> faceting...
> ie. (a) i could facet my emails....then do view a timeline about the
> faceted onces...this would be great way to visually pinpoint that
> email about that meeting two weeks ago....
> and (b) i could timeline my emails...then facet....we were discussing
> something important via a month ago.....the facet by tag (work /
> project x)...
I've been working on such interfaces but for other kinds of information.
See the examples here
http://simile.mit.edu/exhibit/
> Interestinig ideas....(and I can't wait for you to do it ...so I get
> the side benefit of making my life easier ;-)).
:) My pleasure.
David
Ron K.
Tb 3 is aiming for a fourth quarter 2008 release with some Alpha and
Beta milestone releases. The roadmap is the place to see what the goals
are.
--
Ron K.
Who is General Failure, and why is he searching my HDD?
Kernel Restore reported BSOD use by Major Error to msg the enemy!