[Foundation-l] English Wikipedia ethnocentric policy affects other communities

Neil Harris usenet at tonal.clara.co.uk
Thu Dec 21 03:12:47 UTC 2006


Gerard Meijssen wrote:
> Neil Harris schreef:
>   
>> Gerard Meijssen wrote:
>>   
>>     
>>> Neil Harris schreef:
>>>   
>>>     
>>>       
>>>> Birgitte SB wrote:
>>>>   
>>>>     
>>>>       
>>>>         
>>>>> en.WP is not the primary Wikipedia for many wikimedia
>>>>> editors.  These usernames *are* in the primary
>>>>> character set of the users home wiki.  I think it is
>>>>> perfectly reasonable to expect to not be blocked just
>>>>> because Wikimedia adopts SUL.  Bascially what this
>>>>> boils down to is someone who today edits as an IP
>>>>> without harrassment will be blocked after SUL as soon
>>>>> as an en.WP admin spots them on RC.  This will happen
>>>>> without the editor doing anything differently on their
>>>>> end.  
>>>>>
>>>>> There is strong support for SUL throughout all
>>>>> communtities, but I cannot support going live with it
>>>>> if large chunks of the wikimedia community will be
>>>>> locally blocked without warning on any particular
>>>>> wiki.  If en.WP does not wish to adjust thier policies
>>>>> to this mew development perhaps en.WP can be left out
>>>>> of SUL.  Is it possible for en.WP to keep their own
>>>>> user database while all other wikis use a shared
>>>>> database?
>>>>>
>>>>> Birgitte SB
>>>>>
>>>>>   
>>>>>     
>>>>>       
>>>>>         
>>>>>           
>>>> Absolutely.
>>>>
>>>> Different-script blocking on en: must therefore stop as soon as SUL is 
>>>> implemented.
>>>>
>>>> But, since human-readable usernames are essential to managing the wiki 
>>>> system, and for many people names not in their native script are not 
>>>> human-readable, there must necessarily be some sort of solution to the 
>>>> name-incomprehensibility problem before SUL goes live.
>>>>
>>>> The truename + nickname idea looks to me like the favourite at the 
>>>> moment. I definitely support it, with the provisos that:
>>>>
>>>> * truenames _and_ nicknames need to be made globally unique within the 
>>>> single SUL namespace, otherwise they will present new opportunities for 
>>>> misdirection (eg. nick of one user is the same as the truename of 
>>>> another and vice versa, or that on another different-script wiki...)
>>>> * that each truename or nickname have characters only from a single 
>>>> script system (although they will of course in general have different 
>>>> script systems from one another, because that's the whole point)
>>>> * that the anti-spoofing code be used to filter both (which may need 
>>>> some overdue improvements in the anti-spoofing algorithm to be put in 
>>>> place, which I've got on the back burner already anyway)
>>>>
>>>> I would also suggest, at the risk of further controversy, that having an 
>>>> appropriate nick be made compulsory if you intend to use your account to 
>>>> edit in a wiki with a default script different from that of your username.
>>>>
>>>> Perhaps we could put in place some sort of human-driven wiki-based 
>>>> please-translate-my-name-into-your-script service for this, or simply 
>>>> generate a (also globally unique) machine-generated ID until users 
>>>> create an appropriate nick, with regular nag messages to remind them to 
>>>> do so?
>>>>
>>>> -- Neil
>>>>     
>>>>       
>>>>         
>>> Hoi,
>>> You cannot have the cake and eat it too. All languages and scripts that 
>>> we have are human readable. The notion that they are not is brain dead. 
>>> The idea that nicks have to be unique too will be next to impossible to 
>>> implement. Again, when a user is created in a Latin script, and he does 
>>> "evil" things, you will have to check the talk page / history to find in 
>>> what way you want to impose sanctions. This is no different from people 
>>> with a user in another script.
>>>
>>> Again, I think that the suggestion of forcing to have people change 
>>> their name is the wrong way to go. I am dead against it. It is 
>>> discrimination pure and simple.
>>>
>>> Thanks,
>>>     GerardM
>>>
>>>
>>>   
>>>     
>>>       
>> "Human readable" != "readable by all humans". For example, most en: 
>> editors simply cannot read most, or in some cases any, non-Latin 
>> scripts. That is to say, things like "मुखपृष्ठ" and "सुस्वागतम्", or 
>> "𐌷𐌰𐌿𐌱𐌹𐌳" and "𐌰𐍃𐌴𐌹𐌳𐍉" and "கிபீடியாவில", are as 
>> incomprehensible to them, and as difficult to memorize or tell apart, as 
>> "£$^&(*%^$&" and "$($"£%^&$&". And that to a reader of any of the 
>> scripts in the example above, the reverse may well also apply.
>>
>> Faced with a vandal called "सुवाग", most en: users won't be able to 
>> distinguish them from "ागतम्". without looking very, very carefully at 
>> the respective rendered character strings: and often, not even then. Nor 
>> will they be able to think of those users by their names: if anything, 
>> they will think of their names as unintelligible squiggles, and hence 
>> their users as anonymous -- for if you can't understand or recognize a 
>> name other than as a squiggle, what use is it to you?
>>
>> This will be very damaging to the fundamental social fabric of 
>> Wikipedia, and lead to an us-and-them mentality.
>>
>> On the other hand, if our fictional user 武 has his name rendered beside 
>> a Latin script nick on en:, other users will instantly be able to put a 
>> name to him: specifically, his real name, Takeshi, the same one as in 
>> the kanji version of his true username, which will always be present 
>> beside it on en: (Indeed, if they talk with or about him for long 
>> enough, they will come to recognise that kanji character, too.) Of 
>> course, on ar: he would have a different nick: which would also be 
>> either another transliteration of "Takeshi", or some close variant.
>>
>> And the same would apply symmetrically in all of the wikis: if I want to 
>> edit in ar: I should either ask someone nicely to translate my name as 
>> an Arab nickname, or put up with whatever Arabic-readable nick the 
>> system automatically generated for me in lieu of a proper nickname.
>>
>> So: to the technical requirements in my previous E-mails, I'd like to 
>> add a couple of other, social, suggestions:
>>
>> * it should be a guideline that users should always attempt to make 
>> their nicks the same as their username, or a close variant if that is 
>> not directly possible;
>>
>> * and it should be frowned upon _not_ to create a nick if one is editing 
>> for any significant time on a foreign-script Wikipedia (that is, 
>> sticking with a machine-generated nick like x105234 should be regarded 
>> as antisocial). Ideally, the system would generate periodic nag messages 
>> to remind users to do so.
>>
>> I hope you can see that in this scheme:
>>
>> * no-one is asking _anyone_ to change their name
>> * all script systems are treated equally, without any privileged 
>> character set or language
>> * everyone is actually more able to see, read and identify with other 
>> people's real names than if they were presented on their own
>>
>> -- Neil
>>     
> Neil,
> I have some 170*2 users on the Wiktionary projects all called GerardM 
> and RobotGMwikt. With SUL people will be able with one click to see that 
> this is indeed the same person on all of these projects. RobotGMwikt has 
> probably the most edits on most Wiktionaries. The way things are 
> configured are such that I can only have one user. I can promise you 
> that I will not change any of these users. If I am forced to do this, I 
> will stop running the bot that I have been running for more than a year 
> without any complaints as to the naming of these users.
>
> I will again say how much I resent this policy of the English language 
> Wikipedia, it is discrimination pure and simple. With SUL it will be 
> easy to check WHO it is that is doing the editing. The most charitable 
> way I can look on this huha is that when SUL is finally there, this 
> discussion will prove to be a storm in a tea cup.
> Thanks,
>      GerardM
>
>   
Which part of "no-one is asking _anyone_ to change their name" did you 
miss in the above?

Under any of the most recent proposals aired here, you won't have to do 
a thing. Some kind of default disambiguation tag will simply be 
displayed next to your username in all the relevant contexts: probably a 
number in Arabic numerals.

And voila! All your accounts, robot and human respectively, are 
disambiguated for those who can't read the Latin alphabet, in the same 
way as the hypothetical User:武 (102343), would be, before he decided 
(for purely social, human reasons) to change his nick, to be displayed 
as User:武 (Takeshi) on Latin-script wikis, without changing his true 
username at all: it was, is, and always will be just "武", regardless of 
how many wikis he uses it on.

Thus, the only requirement for all users would be the need to recognize 
strings made from the symbols "0 1 2 3 4 5 6 7 8 9", which is about as 
culture neutral as it is possible to get, as people from most cultures 
already either use, or can recognize the digits of the Arabic number system.

And users who don't want to do anything won't have to do anything at all.

If this does not satisfy you, perhaps you can come up with a better 
solution which will let Latin-only readers distinguish at a glance 
between 漢, 源, and 装, (only one of whom is a vandal) and make them 
sufficiently memorable that they can remember them distinctly some time 
later?

-- Neil




More information about the foundation-l mailing list