Tim32

/ mini-overhaul ^

Posted On Monday 19 November 2012 by Murray Colpman (Muzer)

We have forever been noticing constantly issues with /, but never getting around to fixing them. Now, we finally have. In order from easiest to hardest:

* Nomenclature. The two characters together of an opening thing, for example /*, is called a marker. The / is called the mark, and the second character is called the er. Note that */, for instance, is called an ermark!

* Pronunciation. We realised that there isn't actually a nice place to find the pronunciation for the ers. Initially I tossed around the idea of changing the ers to be the actual character pronounced (c, p, k, m, etc.), but neither of us liked this idea, so we decided to keep it as it is and just add the er characters to fonti. This will be done shortly.

* Nesting and escaping. Nesting is now allowed, but it'll probably take an age for it to be added to VOSFace and even longer for TCP. Escaping has always been a problem we've ignored, but not for longer! Escaping in text and in speech works in a similar way but may be done at different points due to the obvious differences between pronunciation and writing in this special case. Only the mark character or sound, if it appears in text, needs to be escaped; the ers never do. To escape the mark, you simply write or say it twice - for instance, /*This is a website tim32.org//blog*/, or (in speaking) /"I want a bi^it of toast"/ (where ^ is the glottal stop, of course, as in usual fonti). Of course, this means if you do say a mark sound and want it actually to be interpreted as a mark, the next sound you say cannot also be a mark - this shouldn't be a problem, as this would rarely occur, and if it does, in the last post you saw that a schwa can always be added to avoid ambiguity!

* We've overhauled !. It now doesn't mean a proper noun, and in fact can never be used for foreign text. You'll see what the replacement is later. ! is now used for Timlan text that is not a Timlan word. There may be some ambiguity as to what this means, so I'll spell it out:
** Timlan text is anything that is pronounced with fonti.
** Not a timlan word means that it is either a special case construction (like eli-) or nothing to do with a word at all. This does not extend to numbers as they are so vitally important. Examples of things that are timlan text but not timlan words could include spelling something out with the Fontibet (/!m'uz'er'!/ for instance), and elements and whatever we later decide to use for compounds.
This is so that parsers won't have to get out of hand coding for special cases so that parsing doesn't break if they come across a special case they don't even care about, but they can still parse the special case if they do. Elements are a prime example of this - now a parser can simply ignore /!elih!/ if it doesn't care about elements, and still be able to parse the timlan properly. Finally, timlan proper nouns *not* in the dictionary would also fall into this category, but we're not sure whether those should or shouldn't exist yet. If we decide they do, they'll be covered by this.

* We've added a new er. The new er is :, and is finally an er for specifying foreign text! It is pronounced as timlan "x", for "eXternal". "But wait", you might ask, "how do I specify the language in which the foreign text is written?". Well, the answer is simple - use /!!/ with the ISO-639-3 code! ISO-639-3 code is a standard that aims to define a code for every significant known language ever in existence, so you shouldn't have too many problems with missing languages, with one obvious exception - Timlan isn't there! If we ever have a need for a code, we'll figure out something else, but for now, this should never be needed for Timlan phrases anyway. You specify the code by spelling it out inside the /!!/ marker/ermark - for instance, /!eng!/ (pronounced /!en'g'!/). You attach this to the /::/ marker/ermark by using the inilan verb or the alan suffix (these will be added to the dictionary/fixionary in due course), thus: relainafelem /!eng!/alan /:France:/ eateht hat~s ei. If you do want to specify a very exotic language that doesn't have a code, but does have a name in a language that does have a code, you can always chain suffixes together thus: haiseht /!eng!/alan /:Na'vi:/alan /:Palki:/ eiami cute