Talk:Sugar design review 2
The graphics are very stylish and have an appealing colour scheme,
In the picture Sugar-private-chat-mockup.png it says the following.
Oh, I forgot to mention that we need names for those characters, if you can think of anything.
Could you please say for which characters you need names please?
- A link from the news of 2006-06-11 may provide the answer!
 Suggestions for names for the characters
Crane (named after the bird, so localizable everywhere)
Swan (named after the bird, so localizable everywhere)
- These names are not localizable everywhere for 3 reasons:
- In many languages, these names will be either feminine or masculine which conflicts with the gender of the Child. For instance the Russian word for Swan is masculine gender.
- In many cultures, specific characteristics are associated with specific animals. The animal names that you suggest, or other animal names, may have negative characteristics in some cultures.
- Cranes and Swans are not necessarily found in every ecological niche on the planet. They may be totally unknown to some kids. For that matter, the names might both translate to the same word meaning big-foreign-bird
- May I suggest Japanese names since the OLPC will likely never be deployed in Japan. Therefore these will appear to be foreign-sounding or Japanese-sounding names. Japanese girls names usually end in -ko so something like Miko and Yoshi would work. Or Gecko and Sushi...
In Presence.png, the Mockup of what the presence pane might look like, what are the two graphics with the arrows in the Everyone section intended to symbolize please?
I thought the laptop will be used primarily in black & white. I guess that is not considered important at this mockup stage?
From Alan Kay Two things to really think hard about:
a - tiny screen real estate b - need for children to construct with images and programs
(a) A good ploy would be to have the first UI that the kid sees have lots of obvious buttons to push. But when the kids start to make their own content, especially dynamic content, they will want every pixel (and the designers need to figure out how to make them available). This is a tricky design task but very necessary.
(b) communication and community are good things, but two of the most important things that the children need to learn (regardless of whether their governments think so) are mathematics and science. The children need to be able to construct simulations of all kinds of ideas and speculations. Both the simulations and the UIs used to help construct them take a lot of screen real estate. Doing this well is probably the most difficult and important UI design task for the HDLT.
UI is User Interface
- If you have a look at Squeak's UI, you see that it does deal with the issue of letting the user get at the whole screen for their own UI. The same is true of most game programs even those running under Windows.
HDLT means ?
- This is Alan Kay's personal acronym meaning Hundred Dollar Lap Top. So far it hasn't caught on elsewhere.
Do you think that 600mhz cpu with slow gpu can make it?
- Strange question. Are you aware that this is an educational project and not a laptop project?
Yes, but some things just can't work on slow cpu.
What about combination of SWN and one level wireless network, a lot of childrens can't acces internet normaly in countries with dead infrastructure or in countryside , but if we could create this net it will open great opportunities for these childrens(instant wireless chat, "locator", vcard exchange, file share...) . Sorry for my bad knowlegde of english lang.
 Anything to learn from Groupkit?
Have you looked at Groupkit to see if you can make use of any of their backend code? They implemented something called Teamwave with their groupware comms library that had similar features to what you describe as "shared documents".
You don't explicitly say it, but when you described the type of documents that would be shared, it sounded like XML/SVG/HTML. Is this correct?
 How to find shared documents without URIs
Here is an idea for naming and finding shared documents. First of all, the names should be Unicode strings, i.e. matching Python's "\w" Unicode regular expression. There is no other structure to the name.
Let's start by assuming that there are many possible namespaces in which the name could exist. On the user's laptop, there will their own namespace. Each buddy also has a namespace. Some of these buddies may have published a portion of their namespace so that you can find the name, even if they are not currently in wifi range. And then there are the special buddies known as teachers, librarians or just plain servers.
If you search these namespaces successively, rather like the way Hypercard searches its message path, then you can either pick the first hit, or provide the user with a list of possible hits.
Rather than viewing this as simply searching folders for filenames, it is better to explicitly recognize that these are namespaces. At some point in time, there will be a need to name things that are not simply files. For instance, a buddy name would refer to someone's identity information in the buddy list that you want to share with a friend. Then, when that friend travels to the next village, this buddy will pop up on their laptop when they are in range. Similarly, applications could be named, and thus shared, even though an application may reside in many files. Or database records may be named so that a record can be shared without sharing the entire databse. For instance, an appointments database record or an email message.
The above is a good idea.
Would it also be a good idea to say that each file name consists of two Unicode strings, namely the name itself and an accompanying tag string. Then everything would work as suggested above and in addition the tags could be displayed as well, and processed in some circumstances.
- Are you suggesting that names should have attributes other than simply the content?
- No. However, if you think that that should be considered.
- This drifts towards RSS which is not necessarily a bad thing since RSS is for distributing a small amount of information about a document that you can retrieve later. If you want to pursue this thread I would suggest that ATOM is more up to date as a model.
- I shall try to follow up that link as part of my own learning, yet the two strings would each be just ordinary strings.
- I notice that it is XML based. I have a problem with XML because it not only uses the < and > signs for other than their basic Unicode meanings but also then causes problems for non-XML text entering an XML-aware system where such text uses a < character as a normal Unicode character. I have raised this before elsewhere yet appear to be in a minority of one! Yet, so! It appears to be a legacy from 7-bit ascii days: I feel that two new Unicode characters such as OPEN AN XML BUBBLE and CLOSE AN XML BUBBLE need to be introduced rather than use < and > as not having their basic Unicode meanings.
- Returning to the matter of your idea, which I think is good as it would allow the children to name an item using their own language with spaces in the name as well if they choose and not bother about which directory it is in when telling someone else about it. What we need to do is to work out what happens if another laptop simply receives a Unicode text string. Is that the name of a file which is sought or is it a message or what? There would need to be some way of sending some indication of meaning to the other laptop. This could perhaps be added "behind the stage scenery" so to speak when the Unicode string is sent by a child, say, clicking on a "Please get this document" button.
> Similarly, applications could be named, and thus shared, even though an application may reside in many files.
With the tag facility, the names of the files for an application could all be the same with the tag used for either the name of the individual file or for something like 2 of 17 or for a combination of those two possibilities: the tag idea could also be used for chapters of a book or units of course material.
How would the Unicode string be terminated? Some software systems use a null character. Would a RETURN character be a better idea? Or maybe one of the other Unicode control characters?
- This is handled by the individual language. I believe Python uses no terminator, it simply keeps track of the length.
- Suppose that in time (maybe many years) the software system of some OLPC laptops is different from some that on some others: if the strings are defined in a portable format from the start then many potential problems could be avoided. I suppose that that is not part of the string itself but is the method of telling a receiving OLPC laptop that the string (or pair of strings if you so agree) is a document name for the purpose of finding documents with that name and asking that the receiving computer please have a look for them when it gets the time!
I suggest that your idea needs a new wiki page in which an open and portable format can be defined. That way, it will be ready for use when people need it and, by being ready for use, may well get used.
The naming of the page and starting it is for you as it is your idea that will be in it.
The page is (please replace this line with a link).