Learn Mando'a

Ha. I lold.

Vod, i donno about all that canon/non-canon stuff, thats your area of expertise. Only thing i know is that the font im using i downloaded from the page on Karen's site, which isnt available now. And i downloaded it ages ago.

And as Karen herself stated (somewhere...) Mando'a is a phonetical language. And why on earth should it be just a swap from latin transliteration? Why another language should contain the exact copy of english alphabet? Wouldnt it be weird? Especially thinking that it would be second exact replica, since Basic is english written with funny letters?

Oh believe me, I'm the first in line for a phonetic writing system. I believe that if you can read the word, you should be able to pronounce it with 100% confidence as well. I also think numbering systems should be logical (not like Arabic numerals, more like Chinese 1-5 and 10). The only caveat is that since we do most of our communication over the interwebs, it would be great if the font could be easily typed with a "standard" computer keyboard without involving too many departures from normal typing flow (look at the various Tolkien Elvish computer typeface keymaps if you want an example of what I want to avoid).

My intent here was just to open a conversation about the fonts we have available. As a result of my research, I just started working on reverse-engineering the original Lucasfilm-commissioned Metschan-designed typeface since that one was never released publicly and I think we should have a strictly-canon font available for those who want to use it.

I like some of the decisions made in the Metschan font (and, correspondingly ErikStormtrooper's newer version of the font) -- most especially the glyph for M, as it is visually similar to N, reflecting the similarity of both sounds (both are nasal consonants). I can live without having C and U looking almost-identical, so I'll gladly accept ErikStormtrooper's version even though it explicitly disagrees with the Metschan font: the character Erik used for C is supposed to be something else, but the only source image I have for Metschan's version is not clear as to what the Latin equivalent should be. P, Q, V, X, and Z can work either way, but the Z in Erik's older font is particularly uncreative (it just looks like a Z), whereas the Metschan V and Z at least fit the "soul" of the system better.

I don't really like either version's decisions regarding numerals -- I would love to just completely disregard the number system and come up with something a little bit more logical.

Whoever designs the numerals should keep in mind it's basically base 5

Well well, maybe something is in the air, as I just posted a topic the other day in the canon mando'a forum basically about this exact same thing, hah.

I'm working on a typed romanized Mando'a character set (for those who aren't familiar with the term 'romanized' it means using standard letters like these to represent the sounds).
I'll be using accented letters to denote different sounds, as accented letters are readily available if you swap to the US-International keyboard layout, which is really quick to get used to using (accents are made by typing a punctuation followed by the letter to accent, such as ' then e producing é)

I may also try my hand at making a handwritten version since that appears to be more on-the-topic in this thread.

Vor'e
-A'nu Biss

Put this together today, my stab at writing practical Mando'a script. I tried writing it out and it was actually far easier to write with pen in hand than with finger on touch screen, hah.



I'll post an example of writing a sentence with it by hand when I've got a moment to scribble on paper then take a picture of it without raising suspicions here at work

EDIT: As promised...


It's 3 lines of the same thing as I practiced writing it. It says "The look of the handwriting"

it's still more complex than I'd prefer, but we're getting there

Vlet Hansen wrote:it's still more complex than I'd prefer, but we're getting there

Yeah I'm not completely happy with the ones that require 3 and 4 strokes, even if it is just a handful of them. I think most of those could be simplified further since there isn't any clarity loss by removing some strokes in them. I also drew them with the assumption that most would write by starting the stroke from the top of the symbol. Most of them I kept to a single continuous stroke. It wasn't terribly difficult to write with them after getting a bit used to it.

One problem I am noticing as I sit here hand-writing various letters of the font: Every character is solely based around a primary vertical line, and the other differentiating marks are occasionally too subtle, leading some letters to look quite similar while hand-written.

And actually, despite my preference for the similarity between M and N in the typeface, handwriting M is a pain and it is difficult to keep it looking different enough from N to be clearly distinguished.

I am also noticing how completely dissimilar sounds have some very similar characters; for example, K and R -- R is just K backwards in the typeface, which makes no sense to me from a linguistic perspective, even though K and R are very similar in any Romanised alphabet. V is just upside-down L. C, G, T, and W are just mirrored versions of one another with some light differentiating tick-marks. P is just A with a tail.

Here's what it looks like when I write it: I tend to write a little more angular and translate the end triangles and such into flicks instead. I might blame learning Chinese for that.

You might notice some small changes I made:

• simplified H to only have a single mark over it rather than 3
• simplified M and moved the tick-mark inside the footprint rather than having it outside (because writing a squared-C shape is not comfortable nor fast)
• simplified every triangular stroke into a "flick" as mentioned above

After completing this exercise and looking at it, it's not awful for handwriting.

That said, I think if we're going to go the phonetic direction, we might want to rebuild the typeface from the ground up to better-represent the sounds each character makes, taking a few cues from Tengwar and the IPA. The new typeface should have a similar "soul" to the original, but more variance in shape amongst the characters (it feels weird to have every single character be a vertical line with some various additional marks hanging off them), and sounds which come from the same parts of the mouth should have similar characters (with consistent differences in the characters to represent the differences between the sounds, like Tengwar consonants or Hiragana/Katakana).

I would like to have them fit into a chart like this: and I would love for them to share visual similarities based on sounds like this: The challenge, I think, will lie in meeting all the following criteria:

• expressive
• simple
• accessible to new learners
• interesting for longtime conlangers
• somewhat applicable or adaptable to non-Mando'a languages

To help illustrate my point about why I think the characters should be representative of their sounds, I put the MANDOR font characters in to a chart organised by sound: Hopefully you can see what I mean when I say that very few of the characters share visual similarity with their neighbours in the chart. There are a few notable exceptions:

• the voiceless plosive p and the voiced plosive b (this one in particular brings me great joy)
• the alveolar voiceless fricative s and the alveolar lateral approximant l
• the alveolar voiceless plosive t and the (modified) palato-alveolar voiceless fricative ch (which is a combination of the sound of the alveolar plosive t with the sound of the palato-alveolar fricative sh -- another way to write ch is to write tsh, and the IPA for ch is tʃ, where t is t and ʃ is sh)
• the alveolar voiceless fricative s and the alveolar lateral approximant l
• the palato-alveolar voiceless fricative sh and the palato-alveolar voiced fricative j (which is actually dʒ but I wedged it into ʒ so it would fit in the chart)
• the velar voiced plosive g and the velar approximant w

I think with a few modifications to highlight the difference between voiceless and voiced as the primary focus. As an example of what I mean:

• swap the characters for k and w
• swap the characters for d and ch
• swap the characters for z and l
• make r and l more similar
• make y and w more similar
• maybe make m and n more similar
• maybe make ' and h more similar

Here's an example of the kind of changes I am proposing, using only pre-existing characters (combined from both the Mandor font and the Metschan font) for sake of laziness: Please note this is just an example, and not my finished product. I have to do a *lot* more thinking about the consistency issue. And about what to do with the vowels (I borrowed one or two vowel characters in order to fit my needs).

Wow, very in-depth, but absolutely needed I believe.

Yeah, I think the language would be quite a bit more intuitive if the lettering was influenced by the part of the mouth / sound produced. Something that will be helpful in determining the shape of the Mando'a characters is figuring out which characters are used the most. We have a sample size readily available in the dictionary. If I have some time today to do so I'll try to extract the Mando'a words from it and do a frequency chart. The letters that occur the most often are the letters that are quickly made more efficient to write by handwriting laziness alone over time.

I concur. If you don't have time to do the frequency analysis, I can try writing up a script to automatically analyse, though I'd have to go through the dictionary by hand first anyway in order to extract the individual sounds we'd be representing for use in constructing the search pattern(s).

Completed the frequency analysis based on the dictionary in Taljair's Total Guide.

Sorry that I have to paste it in Image format - I could probably go through the trouble of putting it into table form but the existence of some characters that wouldn't transfer in to Mando'a for the forum font and the tedium required is preventing me from doing this x)

Anyway, that being said, here it is - if anyone wants the excel format of it I can upload that and share it from google docs or something.

Columns from left to right: Mando'a Character | Actual Count | Frequency as a percentage of all characters | Frequency derived by count / wordcount



Edit log:

I'm curious to know how you set up the frequency analysis. If the spreadsheet will show your work, I'd love to see it.

Did you do the frequency analysis of the occurrence of those characters in the wordlist, or of the sounds they represent in the Romanisation of each word? I'm mostly curious as to which vowel sounds and diphthongs need to be represented across the language (ignoring the existing miitgaan) so we know where we need to focus in updating the miitgaan to be fully-representative.

Also, how did you get the wordlist from a PDF into a spreadsheet? Copy-pasta from PDFs in my experience tend to be unreliable at best, since the PDF format focuses primarily on vector shapes and location coordinates and tends to ignore document formatting stuff like tables and spacing.

Adi'karta wrote:Also, how did you get the wordlist from a PDF into a spreadsheet? Copy-pasta from PDFs in my experience tend to be unreliable at best, since the PDF format focuses primarily on vector shapes and location coordinates and tends to ignore document formatting stuff like tables and spacing.

NitroPDF, ner'vod, is amazing. With NitroPDF I opened the Total Guide and was able to bandbox Text-Select only the Mando'a words. When I copied and pasted them into Excel (actually LibreOffice Calc) it recognized the font from my installed fonts and automatically set it to Mandalorian font in excel. NitroPDF is leagues better than Adobe in just about ever form, but with text interpretation in particular it is generations ahead. It helps also that this PDF isn't a collection of scans or photos, it is clear typeface and fully digitally created material.

Once I had all the Mando'a in Excel (a bit tedious as I could only select from one page at a time), I looked around online a bit until I found a good character counter. Pasted the Mando'a into the character counter and let it go - it spit out the results for every character. Had to modify some of them a bit as Beten was represented about 3 different ways in the Mando'a, so I combined those, removed other punctuation (parenthesis, commas, ellipses, etc.) from the list, and put the results into the same Excel document.

Adi'karta wrote:I'm curious to know how you set up the frequency analysis. If the spreadsheet will show your work, I'd love to see it.

Did you do the frequency analysis of the occurrence of those characters in the wordlist, or of the sounds they represent in the Romanisation of each word?

There were 2060 words total. To produce the frequency percentages I totaled the number of occurrences of all characters from the character counter output and just divided each character's count by that total, pretty straight forward. Same for the word percentage frequency count, but dividing the characters' counts by the total number of words (2060).

To answer your question about sounds vs Romanisations, this is a direct character count of the Mando'a script in the Total Guide PDF, so it is the sounds themselves (e.g., 'uy,' 'ee,' 'ch,' 'sh,' etc. are represented by a single character in the character count, so 'ch' and 'sh' would not produce counts for C, S, and H)

Here's the link to the Excel file if you want it, but I don't know that it'll "show my work" as you've requested. Hopefully my explanation above covers that enough? https://drive.google.com/open?id=0B3VsK ... URPaXpFSDA

-A'nu Biss

Could have just asked for the original... I work in Excel myself

Taljair te Mir'ad wrote:Could have just asked for the original... I work in Excel myself

Oh, Hahahaha.

Well yeah that certainly would make things easier. Can you share that with us please? I'm trying to create a Quizlet for learning the vocabulary, and having those word definitions in a format where they can be easily copied out would be a life saver

posted my latest here:
viewtopic.php?f=19&t=108

Adi'karta wrote:One problem I am noticing as I sit here hand-writing various letters of the font: Every character is solely based around a primary vertical line, and the other differentiating marks are occasionally too subtle, leading some letters to look quite similar while hand-written.

That has to have been intentional, the vertical line. I'm sure we could come up with all kind of interesting cultural or developmental reasons that the vertical line is featured so prominently, but, I have to say that I don't mind at all. English has a lot of the same - RTIPDFHJKLBNM, and arguably even WYUAC fit that mold. That said, they are hard to distinguish at a distance or a glance, and I found the easiest way to do this is to completely ignore the vertical line, focusing on the direction and number of other markings.

Adi'karta wrote:I tend to write a little more angular and translate the end triangles and such into flicks instead. I might blame learning Chinese for that.

You might notice some small changes I made:

• simplified H to only have a single mark over it rather than 3
• simplified M and moved the tick-mark inside the footprint rather than having it outside (because writing a squared-C shape is not comfortable nor fast)
• simplified every triangular stroke into a "flick" as mentioned above

After completing this exercise and looking at it, it's not awful for handwriting.

I like pretty much all of these. I handled hand-writing the M a different way - handwriting sample pending.

Adi'karta wrote:That said, I think if we're going to go the phonetic direction, we might want to rebuild the typeface from the ground up to better-represent the sounds each character makes, taking a few cues from Tengwar and the IPA. The new typeface should have a similar "soul" to the original, but more variance in shape amongst the characters (it feels weird to have every single character be a vertical line with some various additional marks hanging off them), and sounds which come from the same parts of the mouth should have similar characters (with consistent differences in the characters to represent the differences between the sounds, like Tengwar consonants or Hiragana/Katakana).

I like the idea, but, unless it's parent letter is immediately and unmistakably apparent, I think this is a bad idea. Someone learning mando'a, seeing this easier type face, should be able to also read anything in "old mandalorian" with only token difficulty discerning the characters. So, not so much a bad idea, as, the absolute priority in my mind would be consistence with canon.

That being said, if you hand-write anything enough, you end up with something way more streamlined than a computerized type-face. I found that after repetition, hand-writing mando'a to learn the symbols, many of them became rounded. There's only really a few variations of the same difference, and its direction and/or angle is what defines it - once you write this fast enough, it becomes its own distinct character.

The best way, before uploading my own handwriting sample, to explain is to point at the lower-case of: b, d, p, q (in english) - the same verticle line with a loop is made for each letter, but they're quite distinct.

Enaris wrote:...There's only really a few variations of the same difference, and its direction and/or angle is what defines it - once you write this fast enough, it becomes its own distinct character.

The best way, before uploading my own handwriting sample, to explain is to point at the lower-case of: b, d, p, q (in english) - the same verticle line with a loop is made for each letter, but they're quite distinct.

My recent exploration in learning Thai has shown me countless examples of this. So many letters in Thai are extremely similar, some only differentiating based on the direction you begin the "head" of the letter (the little circle on it), clockwise or counter-clockwise.

So, that being said, letters that are extremely similar, almost indistinguishably different in a typeface, quickly take on more discernible differences when written repeatedly, and anyone who is intimately familiar with the language will have no problem telling them apart at all, even in the typeface.

Is this conversation still alive? Because I'd love to help in any way I can!