I want you to update on what’s going on in hamster’s master git!
In short, tags start to appear, there is keyboard frendlier autocomplete window, raising dependency of gtk+ to 2.16 to make life easier for translators and, errm, other stuff.
But the tags are the big bit, and i will talk about them. Here’s the main dropdown for you:
Main window featuring tags input field with active drop-down showing matching tags. Tags can be up to two words long and separated by comma.
When I started to look on tags, I did my best to look around and see how others have solved it. In my search I found Freckle. I encourage you to check it out – it’s a web based time tracking solution. Although it is not for free – for some it might even be a better fit than Hamster.
Checking screenshots and the tutorial some might notice that hamster’s auto-complete looks quite similar.
I’ve indeed replicated part of the behaviour and the general idea of a flowing tag list instead of a vertically arranged one. Although there are at least thirty three ways to tie your shoes, i don’t think there is a valid reason to deny something that is working just to be special. The replicated behaviour bit is how they distinguish between a description and a tag. A description is anything that is longer than three words or starts with an exclamation mark. Quite simple and makes perfect sense. Anyway, that’s about where the similarity ends.
Update: We dropped the description-is-anything-longer-than-two-words. Description now again goes into the main input field and tag field is just for tags.
Current status is that you can add and edit tags in activities and you can adjust which ones will appear in auto-complete via preferences.
There is lot that could be done. But for starters it would be nice to get them in the overview and reports, and then think about filtering, searching and a tag view.
If any of frecklers are reading this – hope you don’t see hamster as a threat, I wish you guys luck and keep rocking!
In other news – if you are looking where to exercise your Python skills, i can strongly suggest getting in contact with Og Maciel and help him with BillReminder. A fantastic person and certainly fun to code with!
Of course, you could also contribute to Hamster, you busy busy busy person.
Just leaving a short note for morally support this feature! The more I use hamster, the more I like it, and I like the direction the project is going.
Af for contributing to Hamster: I renew my availability (https://bugzilla.gnome.org/show_bug.cgi?id=596999#c5) to have a go with the documentation of the project, if you give me some input on what is the format / content that you envision. You have my mail, so if you wish you can contact me!
Long live the critter!
Just curious: where does Hamster get the tags from? An internal database or from a desktop-wide repository such as Tracker? The latter would be coolness, of course.
As for lists, the rationale behind vertical arrangement is that it’s faster to find the tag you are looking for when you only need to move your eyesight a few pixels up or down to read the next tag in the list. In a flowing list such as yours, sight must travel much greater distances to process the same number of items.
I would certainly find the flowing list less intuitive to use – but, arguably, it is more aesthetically correct, in that it does not create big amounts of unused white space as a vertical list might.
There is nothing obvious about cooliness in using a desktop wide repository such as tracker to store tags in. Not now at least. If a migration makes sence at a later point, we will consider it.
Since you know about usability aspect of lists, you certainly also know that skimming is performed at speed up to 400-700 words per minute[1].
Other reasoning would be the vast amount of tags that one might introduce. As i’ve seen people having 13 categories already, i’m sure somebody will do it’s best to abuse tags too. The rest should be as comfortable with horizontal layout as they are with vertical one.
Rule of thumb when speaking about usability in general is – don’t argue – test it. So, errrm, how about doing that?
1 – http://en.wikipedia.org/wiki/Reading_%28process%29#Reading_rate
Why would I test it?
if you are having existential doubts, i can suggest reading Seneca. Richard Rorty’s quite nice too.
Peh, maybe i started on the wrong foot.
I am well aware of the benefits of aligning elements vertically, but there are more sides to it. A good example is drop-down versus a pie menu. First might be faster to scan, second though is easier for kinetic memory. With all respect, i am not a huge fan of unasked lessons, so pardon me for snapping.
As for the tracker – i really fail to see what would be so cool in having all tags in single place right now. Would just storing tags in single place really suffice? Or would one self-construct SPARQL queries to deduce new wisdom? So sorry there, for being a little sceptic. I admit that i do have a problem with people shouting “this is cool” and “this is not”, without explaining themselves.
You find it particularly odd that people don’t bother explaining themselves when the attitude you display above is the one they are faced with when proposing something that isn’t the status quo?
i’m sorry i hurt your feelings.
nah, just kidding.
Well? Do you?
sure!
Why? Why would anyone want to contribute on those terms? For the glory of making life easier for a bunch of assholes?
that’s rude.
anything else?
Why would storing tags in a single place not suffice? And just for the record, I didn’t say that other solutions weren’t cool. I’m not sure why you would read that into my post. It may very well be that your solution works best for you. I just naively thought that an innocent exchange of ideas would be the best way to engender progress. Alas – I should have known that for every character not tersely written in C, God kills a kitten.
We cannot “store tags in a single place” as we still need to map tags to our database records and these records are part of Hamster, not Tracker. We could of course introduce some sort of GUIDm generate unique URIs for activities and then feed activities to the Tracker indexing engine but then we have to store even more data on the Hamster side. Also it would mean talking to Tracker each and every time we display or edit any piece of data as in Hamster tags are a first-class citizen and are visible in more places than a “properties” window.
What would you do with activity tags appearing in the tracker? Where would they appear? How would they be used? What would be necessary to make that happen? Would it be just tags, or the activities with them?
I’m not questioning you – just urging to think through . And when you get there, you could repost your initial comment with those suggestions.
Otherwise for me it’s all like “oh guys, you could put hamster on mars! That would be cool (of course)!” – cool how? why of course?!
patrys: Clearly, I don’t understand how Hamster stores activities internally. But would it be that impractical to just reference the tags in Tracker directly? That is – instead of storing a list of references to items in Hamster’s tag database for each activity, Hamster would store a list of URIs denoting tags in Tracker.
Toms: You mean how the tags would be used outside of Hamster? Well, for instance, it would mean that you could create a tag called ‘coolproject’ in Hamster when you initiate your activity, then go on to work on an e-mail in Evolution and seamlessly tag the e-mail with the same tag without having to create it anew, since Evolution would operate on the same tag database. This would be nice just for starters, and wouldn’t require Hamster to use Tracker for anything but tags. But later on, if you decided to track metadata of activities in Tracker as well, it would open even nicer scenarios where one could, say, search for all items in Tracker with the tag ‘coolproject’, and have both the activity in Hamster and the e-mail in Evolution show up among the results, or, in an even more advanced version, search for all mails with the tag ‘coolproject’ sent while not working on ‘coolproject’ in Hamster. It’s usefulness would really be a function of how many applications made their metadata available in this fashion.
KDE already does tags this way, by the way (they do it in Soprano rather than Tracker) – it’s not like its a concept that is completely alien/unheard of.
All kinds of nice interaction are possible but until tracker gets integrated into more desktop apps for us it just means adding more work. We cannot really reference a tag as such a reference would be ephemeral (tags are free to go away once no content Tracker is aware of uses them). For proper integration we’d really need to push activities as documents to Tracker and I’d be more than happy to do it once it represents any value. That’s why Toms said we’re more after making Hamster better and easier to use than making it use cool technologies just for the sake of it.
Well, I guess my problem is that I am still oblivious as to how “cool technologies” are diametrically opposed to “better and easier to use”, but fair enough…
You’re more than welcome to come with a patch if you’re willing to write, test and maintain this part of Hamster. That’s what I meant – for me and probably for Toms too there is too little gain to invest our time in such a feature *right now*. As Tracker gains adoption I’m sure it will eventually end up at the top of our TODO list.
Yes, a pack of cigarettes, please. Why not just converse like a normal human being instead of all this attitude BS?
Would you like fries with that, sir?
Do we really need two words for a tag?
For example, I wrote “Having tea” as a description and it became a tag!
I think you want us to use an exclamation mark to make that a description.. IMHO, it should be other way round. Use the exclamation mark when you really want to make a longer tag. That would be least surprising to users.
Thanks.
the reasoning behind why to have two word tags can range from just an idea to a careful research. i don’t know which was the case for freckle guys, but i believe that the more apps do things in a uniform way, the better it will be at the end.
in your example, first i could ask – why would you bother if it is a tag or description, as the information is there, and second – if the description is so short, maybe it belongs to the realm of tags after all? next time you will be having tea, the description will be faster to type in thanks to the autocomplete!
I thought the tags have a different role to play when compared to the descriptions.. may be searching / grouping by a tag etc.,
Btw, In 2.29.5, I cannot search by multi-word tags (Having tea). The search only works for a single word tags.
there are lots of things you can’t do yet.