Talk:OLPC Human Interface Guidelines/The Laptop Experience/The Journal: Difference between revisions

From OLPC
Jump to navigation Jump to search
(Perhaps notes should be "wiki-pages", so that kids could use linking and other features)
(Automatic Grouping and subgrouping of documents)
Line 4: Line 4:


By this I mean that a meta-data description '''has''' cultural values and idioms attached to it. You can't do a simple translation of the term as it will lose all value. There's a local (even community) sense in which foreign words could be translated or adapted (ie: "to 'boot' a machine" is sometimes translated as "'bootear' una maquina"—technically a barbarism, but used nontheless). There should be some way to control the way in which those tags are generated or used... --[[User:Xavi|Xavi]] 23:28, 6 February 2007 (EST)
By this I mean that a meta-data description '''has''' cultural values and idioms attached to it. You can't do a simple translation of the term as it will lose all value. There's a local (even community) sense in which foreign words could be translated or adapted (ie: "to 'boot' a machine" is sometimes translated as "'bootear' una maquina"—technically a barbarism, but used nontheless). There should be some way to control the way in which those tags are generated or used... --[[User:Xavi|Xavi]] 23:28, 6 February 2007 (EST)

=== Automatic Grouping and subgrouping of documents ===

Epiphany, Gnome's browser, uses tags to organize bookmarks. This allows to use the location bar to find specific bookmarks via their tags, names and descriptions. But it also provides another functionality. It organizes the tags dynamically into groups and subgroups named after the tags. For example, if there are 20 favorites with the tag programming, and some of them have the tag python, etc. would create programming->python, programming->C . I don't know the exact method being used, but it is very good. It basically takes the good information we provide in tags, and turns them into one logical tree.

Tags and meta-data are excellent ways to organize your information, but grouping and hierarchical organization can also be helpfull, both to give an overview of the grouping and sub-grouping, also to better handle large number of documents. It is possible that when this last point is reached, it will become easier to loose information or forget documents, either because you can't remember the exact tag or name of it, or because you've just forgotten about it (and didn't star it or mark it as being important).

And it is with high collections of bookmarks that epiphany's method shines. When you want to search for a specific site, you do that. When you want to browse, you use the favorites menu. I invite you all to use it for a while, even if just a 15 minutes test, manually opening your other browser's bookmarks into epiphany, and tagging each. After having collected a large number of them, you'll see how beautifully it organizes them (maybe even better than you do!).

Since this organization would be done automatically for kids, it would remove the burden of organizing every document on their laptops into a hierarchy (if it's hard for us to maintain logical and consistent hierarchy, imagine for kids).
[[User:HoboPrimate|HoboPrimate]] 19:37, 24 June 2007 (EDT)


== Automatic Backup and Restore ==
== Automatic Backup and Restore ==

Revision as of 23:37, 24 June 2007

Tagging and Meta-Data

Maybe I misread, but it would seem that many of those tags and meta-data descriptions will be generated in some automatic way (particularly the later). If that were the case, there should be some considerations for language and cultural issues.

By this I mean that a meta-data description has cultural values and idioms attached to it. You can't do a simple translation of the term as it will lose all value. There's a local (even community) sense in which foreign words could be translated or adapted (ie: "to 'boot' a machine" is sometimes translated as "'bootear' una maquina"—technically a barbarism, but used nontheless). There should be some way to control the way in which those tags are generated or used... --Xavi 23:28, 6 February 2007 (EST)

Automatic Grouping and subgrouping of documents

Epiphany, Gnome's browser, uses tags to organize bookmarks. This allows to use the location bar to find specific bookmarks via their tags, names and descriptions. But it also provides another functionality. It organizes the tags dynamically into groups and subgroups named after the tags. For example, if there are 20 favorites with the tag programming, and some of them have the tag python, etc. would create programming->python, programming->C . I don't know the exact method being used, but it is very good. It basically takes the good information we provide in tags, and turns them into one logical tree.

Tags and meta-data are excellent ways to organize your information, but grouping and hierarchical organization can also be helpfull, both to give an overview of the grouping and sub-grouping, also to better handle large number of documents. It is possible that when this last point is reached, it will become easier to loose information or forget documents, either because you can't remember the exact tag or name of it, or because you've just forgotten about it (and didn't star it or mark it as being important).

And it is with high collections of bookmarks that epiphany's method shines. When you want to search for a specific site, you do that. When you want to browse, you use the favorites menu. I invite you all to use it for a while, even if just a 15 minutes test, manually opening your other browser's bookmarks into epiphany, and tagging each. After having collected a large number of them, you'll see how beautifully it organizes them (maybe even better than you do!).

Since this organization would be done automatically for kids, it would remove the burden of organizing every document on their laptops into a hierarchy (if it's hard for us to maintain logical and consistent hierarchy, imagine for kids). HoboPrimate 19:37, 24 June 2007 (EDT)

Automatic Backup and Restore

Would it be possible to backup to another laptop? Or use another media (instead of wireless communication with the server) to transfer things to and from the server? For example two siblings, one of them is sick the other goes to school transporting data/homework/etc. of the sick child to and from the school? Similar to Motoman but also active for this kind of 'transparent' backup.

We've had discussions about implementing a "suitcase" or "messenger" activity, which would allow a child to port another's stuff around, essentially acting like email by foot. Both ends would be electronic, but the gap between the non-connected networks would be bridged by physical movement. If this becomes reality, then I'm sure the ""recipient" could be set to "server". - Eben 12:24, 9 February 2007 (EST)

On a side issue, will the child be capable of explicitely deleting (personal) stuff stored in the server? We've all written/drawn things we don't want around... not even in a backup... ;) --Xavi 12:37, 8 February 2007 (EST)

This is a good point, and I'm not sure the exact answer yet. What I do know is that everything the child makes will be assigned a scope - kind of like privileges - that determine who can see it. This can be private to themselves, open to all of their friends, any private group they are a part of, their class, or anyone. This will add security to the backup, even if it's not removable. We'll see. - Eben 12:24, 9 February 2007 (EST)
Ya know, this sort of indicates two different things that could be important. Sometimes I have something that's important to me or my family or teacher. I want my sister to get my homework to my teacher, and dad want's his letter to get to the other village. But after themessage has gotten to the person who matters, I don't want the messangers to open the left over note. So if it's important, I want it to be a priority, and maybe it's a priority for a period after it's due, but then it becomes less important. Basically, we are talking about either adding explicit importance tags (which burdens the user), or else building an undocumented butler who cleans for you, without asking... Another thing here is what if activity partners place different privacy emphasis on something? This all shows a great need for careful analysis. Is there a set of standard terms to help clarify component or object meanings? --Jeff 00:56, 12 February 2007 (EST)
We're basing much of the security model and also the privacy model on the "chain of trust" idea. With respect to privacy, this means that what you state is certainly a possibility: another participant of an activity I was in might share it with someone outside the group; the friend I gave my drawing to might just show it to someone else I don't know. Of course, this isn't a new problem; this is a problem that every child has faced growing up in school, when notes and photos and other physical media got passed amongst peers, with or without the consent of others. We are certainly emphasizing community on the laptops, but we're not trying to regulate or restrain the community. Many of these problems will be left to the social structures that the children build, since we see this as an important set of skills to learn as well. -Eben 13:26, 13 February 2007 (EST)

Additionally, what about taking advantage of relatively free online data storage services. Google, Yahoo, and other big portal companies offer rather larger repository serives. The network model means that most children will eventually be able to link to the big world wide web through chained paths, so a model of periodic large backup would be very nice. --Jeff

This is already in the works. It wasn't included as part of the initial description because its still in early development and we're not sure exactly how it's going to function yet. The school server will still be the first step, since it will give kids access to everything at least from school, but eventually even the server will reach space constraints. - Eben 12:24, 9 February 2007 (EST)


Eben, as I think about this, I'm sort of wondering if networked mesh drives could become part of the journaling space solution... You're working, and while you are connected, you have networked disk space. So use it. When the mesh goes away, you start filling up your local drives... but if you get that mesh back, it starts moving journal entries to your mesh drives. So we need to keep track of things that are local, and which local things have been meshed. if it's old, but local, it's unique and should be erased. If it's old, and meshed, you'll eventually have access to it again. if you aren't using it today, it can be let go. And some of these servers do have space constraints, but gmail's constraints are so high that it feels unlimited. --Jeff 12:33, 22 2007 (EST)

Personal Information Manager (PIM)

OK, so this all implies that we're thinking of building a custom activity centered around journaling and journal management. Other activities will end up creating memories/notes in the journal that could include clips (forked from the clipboard), Logs (captured from activities), and playmates (captured from shared activities). This entire thing feels a bit like microsoft's outlook. You plan a meeting, share files in the meeting notice, it's email, attachments, and a calender adjustment with a list of attendees. But this implies sugar will go the extra step and take meeting notes for you, since there is also chat information. --Jeff 01:02, 12 February 2007 (EST)

Dependency Handling

Suppose that a student is preparing a report and includes a photograph or two within the text. To avoid copying the photographs, they are instead included by reference in the document. Later, the disk fills, and the Journal suggests deleting these large photographs to reclaim space. The student, thinking they are no longer needed, deletes them, thus breaking the references in the report document. This situation will also occur with video editing, image editing, and potentially many other activities.

To avoid breaking dependencies, the metadata for each file should note the files on which it depends. A dependency should never be suggested for deletion unless all files that depend on it will also be deleted, and users should be warned if they explicitly attempt to delete such files. For efficiency reasons, it may also be appropriate to maintain "reverse dependency" pointers. None of this should be user-visible unless they attempt to delete a dependency. Also, when a student sends an item to another student (or to a teacher), all of its dependencies should be sent with it automatically. Ben 21:44, 11 April 2007 (EDT)

Notes

The text says that Notes will be text based only, if I read it correctly. I understand that these Notes are a sort of Diary + Blog for specific entries kids may want to place public.

If so, wouldn't it be better to use instead the simpler wiki engine being developed by OLPC for Notes? This would allow kids to "journal" their events with more than just text, they could embed images, videos (through gnash or a Player plugin maybe?), sounds, eToys (plugin?), and to refer using links to other notes, documents (the users, but also of friends), and websites?

Even if, for simplicity sake, many features of wiki where hidden in the notes, I think that the excellent features of hypertext and wiki should at least be present there, since not only it would give kids the full power of it right from the start, it would also also allow them to easily move to create wiki books, and to create network wikis.

Is the OLPC wiki engine up to this? Even if not using the more advanced features (like embeding music, videos and sound)?

Just my thoughts. HoboPrimate 18:45, 24 June 2007 (EDT)