Roger Attrill

 UX Chat

User Experience (ux.stackexchange.com) community chat room: ta...
Dec 17, 2015 20:29
hello :)
Dec 16, 2015 18:17
:) - bye!
Dec 16, 2015 18:15
Actually - I haven't given anything away here - it's almost been TOO generic a chat :)
Dec 16, 2015 18:15
but I might be able to make some more generic question out of it
Dec 16, 2015 18:14
hah well I kept it in chat because I didn't really want to make my queries too public
Dec 16, 2015 18:13
@plainclothes thanks for the chat - have to go now but will continue to mull things over.
Dec 16, 2015 18:09
I know what ETL is but you're starting to talk a different language :)
Dec 16, 2015 18:06
It's a 'big data' scenario so results aren't going to be particularly quickly forthcoming :)
Dec 16, 2015 18:06
Thanks - autocomplete is a different thing, but yes timing is still important here.
Dec 16, 2015 18:00
it's the least offensive behaviour and puts user in control
Dec 16, 2015 17:59
yes - that would be the default I think
Dec 16, 2015 17:59
So I was looking for examples to see how they 'felt'
Dec 16, 2015 17:59
and of course the user may just want to wait for all results to come in
Dec 16, 2015 17:59
another is to examine interesting results for a few minutes until more data comes in
Dec 16, 2015 17:58
There's different ways people may use such a search - one is to look for at least one type of result in which case they may want auto update rather than a manual refresh
Dec 16, 2015 17:57
and the order of the databases being searched
Dec 16, 2015 17:57
and the experience may vary hugely for different searches
Dec 16, 2015 17:56
yes
Dec 16, 2015 17:56
There may also be some charts of the data - so not just a list of results.
Dec 16, 2015 17:55
@plainclothes well I suppose the main difference is that Twitter is a linear stream, whereas the results of a federated search may completely override what has been shown already in preference for some more relevant results that have been found.
Dec 16, 2015 08:08
Science.gov is the best example I can find.
Dec 16, 2015 08:07
There are some parallels with feeds like Twitter - the buffering and notification of new updates for example, so it's not entirely unrelated :)
Dec 16, 2015 08:06
The closest 'common' scenario I can think of is email coming into an inbox that you have sorted based on a search term rather than by date, but imagine you're getting millions of emails each time and showing the top 10 or top 50 best results. You wouldn't want the view to change under your feet, so there'd need to be some buffering until the user wants to update with the latest.
Dec 16, 2015 08:01
You might be able to refresh the results at any time or get them to update automatically if you're not interacting with the resulting list of documents but just want to monitor the results.
Dec 16, 2015 08:00
@plainclothes - no I'm looking for more like where the data comes from many sources (that's the federated search bit), but the results are ordered by say for example relevance to your search term, so you'd get a result after a couple of seconds as 'something to look at', meanwhile the search continues in the background and may take a long time depending on the sources and amount of data or complexity of the query.
Dec 13, 2015 19:25
I'm looking for examples of displaying incrementally updated search results such as those from federated search engines (aka meta-search or library portals). The only examples I can find online so far are from Deep Web Technologies who provide the tech behind sites like Science.gov
Anybody know of anything else similar to this, where you get an initial result and can update the view with more results while the search continues in the background?
Jun 17, 2013 14:38
Jun 17, 2013 14:34
@rk yeah - the piano guys !?
Jun 6, 2013 09:31
000 - > WTF
Jun 6, 2013 09:28
404 Example :- masswerk.at/404
May 31, 2013 15:06
@DanHulme Heh - no you're fine! Scotland.
May 31, 2013 09:35
I narrowly avoided my own holiday at a similar time! :)
May 31, 2013 09:28
@DanHulme I should be speaking to the Cambridge Mobile App Group about user experience for mobile on 16 July
May 31, 2013 08:11
A couple of past topics have come up on bbc news recently. Apologies anyone outside the uk who can't read these:
Pareidolia: http://www.bbc.co.uk/news/magazine-22686500
The secret button at road crossings: http://www.bbc.co.uk/news/blogs-ouch-22706881
May 13, 2013 10:21
@DanHulme well if nothing else you've reminded me how useful it is to have an IDE with good productivity tools built in. I use MSVS C++ a fair bit. I used to have Whole Tomato's Visual Assist X in a previous job, so I've just gone to get it again.
May 13, 2013 10:10
Well - thanks for the pointer - I've not seen quite such a tightly integrated search solution before!
May 13, 2013 10:09
That's very powerful - something designed into the framework from the ground up - like undo/redo
May 13, 2013 10:09
and the string you started to search for (before you selected the settings dialog) - you can continue typing in immediately in the settings dialog itself.
May 13, 2013 10:05
straight to the 'show line numbers' setting - nice!
May 13, 2013 10:04
A single point of access to all actions and commands
May 13, 2013 10:04
holy crap!
May 13, 2013 10:03
haha
May 13, 2013 10:03
'jinx'
May 13, 2013 10:03
So now you expect it everywhere!
May 13, 2013 10:02
And as you say - great for long lists/trees - where you know something is there
May 13, 2013 10:02
It feels good to use
May 13, 2013 10:01
The discoverability might be a small issue, but it's also an 'expert feature' rather than something that needs to be obvious before you can even use the app.
May 13, 2013 10:01
@DanHulme hmm - that's quite nice - on single select trees you just type, then up down cursor keys change to prev/next match, return to activate - I've made a very similar interaction on trees for someone a couple of years ago, but the difference here that I like is the 'just start typing' aspect - no need to Ctrl-F (whatever), or hit a search button.
May 13, 2013 09:40
@DanHulme ok - presumably not 'as you type'?