From sinhala-admin@linux.lk Tue Mar 01 21:45:04 2005 Return-path: Envelope-to: lurker@linux.lk Delivery-date: Tue, 01 Mar 2005 21:45:04 +0600 Received: from localhost ([127.0.0.1] helo=penguin.lug.lk) by penguin.lug.lk with esmtp (Exim 3.35 #1 (Debian)) id 1D69YV-0008QR-00; Tue, 01 Mar 2005 21:45:03 +0600 Received: from hantana.pdn.ac.lk ([192.248.40.1]) by penguin.lug.lk with esmtp (Exim 3.35 #1 (Debian)) id 1D69Xr-0008Pd-00 for ; Tue, 01 Mar 2005 21:44:24 +0600 Received: from tissa.learn.ac.lk (tissa.learn.ac.lk [192.248.1.164]) by hantana.pdn.ac.lk (8.12.10/8.12.9) with ESMTP id j21FiMcn095023 for ; Tue, 1 Mar 2005 21:44:22 +0600 (LKT) Received: from localhost (localhost [127.0.0.1]) by tissa.learn.ac.lk (Postfix) with ESMTP id 0413533EA1 for ; Tue, 1 Mar 2005 21:44:23 +0600 (LKT) Received: from tissa.learn.ac.lk ([127.0.0.1]) by localhost (tissa.learn.ac.lk [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 22089-06-2 for ; Tue, 1 Mar 2005 21:44:15 +0600 (LKT) Received: from mail.netspace.net.au (thunder.netspace.net.au [203.10.110.71]) by tissa.learn.ac.lk (Postfix) with ESMTP id 351FD33E6B for ; Tue, 1 Mar 2005 21:44:14 +0600 (LKT) Received: from B1.HOME (220-253-57-171.VIC.netspace.net.au [220.253.57.171]) by mail.netspace.net.au (Postfix) with ESMTP id 0F21643567; Wed, 2 Mar 2005 02:44:04 +1100 (EST) Subject: Re: [sinhala] Re: Sinhala GNU/Linux From: Harshula To: Harsha Senanayake Cc: sinhala@linux.lk In-Reply-To: References: Content-Type: text/plain Message-Id: <1109692077.4600.59.camel@B1.HOME> Mime-Version: 1.0 X-Mailer: Evolution 2.0.2 (2.0.2-3) Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new Sender: sinhala-admin@linux.lk Errors-To: sinhala-admin@linux.lk X-BeenThere: sinhala@linux.lk X-Mailman-Version: 2.0.11 Precedence: bulk List-Help: List-Post: List-Subscribe: , List-Id: List-Unsubscribe: , List-Archive: Date: Wed, 02 Mar 2005 02:47:56 +1100 On Mon, 2005-02-28 at 10:11 +0600, Harsha Senanayake wrote: > > Before you simply categorize Zero Width Joiner as an 'escape character', > > note that it is, as its name indicates, an entirely appropriate > > codepoint for 'joining' two glyphs together. That is precisely why it's > > called a zero width joiner! > > You can call it whatever you want :-) but its a control character used to > enforcing formatting by the rendering software. It doesn't have any > significant value for the text you are encoding, but you encode for the > sake of enforcing how its going to be displayed. I think this is where our viewpoints diverge. I don't see ZWJ as a character for "enforcing formatting", as you do. I simply see it as a mechanism for *uniquely* encoding glyphs such as conjuncts and touching letters whilst retaining the purely phonetic nature of the standard. > > something that parses it. With your proposal how do you use Sinhala > > in /etc/passwd for a username, for example? > > A name with a ligature, 'ksha' has to be written with a zwj now. If the > authenticator does a one-to-one matching of characters, and if the user > enters a key sequence without ZWJ it'll get rejected. Ofcourse, this wont > happen if the authenticator is unicode aware and normalizes text before It's far simpler than this. If ksha needs to be typed, the user would type kayanna + join + shayanna. However, the point I was making was w.r.t. rendering and NOT input. > comparison. Markup is optional it's only used to help the rendering > application pickup correct opentype features. With your metadata encoding scheme, how would you display the usernames from /etc/passwd on a graphical login page where you select the username from a drop-down menu? Do you display 1) Al-lakuna and use two glyphs, 2) Conjunct or 3) Touching? How would you make that decision when there is no metadata? If you propose to include the metadata in files such as /etc/passwd, then your scheme requires all the major libraries to be changed. Which is far too disruptive. > > Education & History! hehe. Simple question. how did you come to know > > that all three forms were equivalent? Were you born with that knowledge > > or did you see it in a *book*? :-) > > Common.. I was not trying to be a smartarse. :-) I wanted to get > enlightened myself. From a ignorant layman's perspective, it didnt make > sense to me coz if it rendered and pronounced the same, without a hex > editor even a *scholar* wouldnt know the difference. But if the touching > letters do not have the 'al-lakuna' and you just need them to render as > just touching letter, then it would have made sense. The 3 alternate forms are pronounced the same but are, obviously, rendered differently. The rendering of touching and conjunct letters do NOT display the al-lakuna. But when encoding, the al-lakuna needs to be there for sorting. > I am not trying to undermine the effort ICTA has put in coming out the > SLS1134, nor am I trying to criticize it.. even though I feel they should > have done much more in i18n/l10n of OSS projects like what other Asian > governments have done (Indix, m17n..) without kissing M$ arse ;-) I cannot > really say SLS1134 is wrong since its serving its purpose and my solution > is just an alternative, a different way of thinking (in FOSS world, we > always encourage alternatives right?.. even though how stupid they may > sound :-) .. May be I am just obsessed with separation of concerns :-) IOC, > DAO patters, XML/XSL/CSS ... I guess SL doesn't really have as many resources as some other countries. cya, # _______________________________________________ Sinhala mailing list Sinhala@linux.lk https://secure.linux.lk/mailman/listinfo/sinhala