Natural Arabetic Input Method (NAIM)

Saad Abulhab's picture

Our site, arabetics.com, was recently upgraded to introduce more options for Arabic and Arabetic scripts users. In addition to adding more fonts and the ability to edit and display extended Arabic scripts online, using any of our fonts, we have also introduced on the site a java applet demonstrating our new alternative Arabetic editing method: NAIM

This new input method option is a “shaping” free (or glyph substitutions free) natural Arabetic text input method. User sees no glyph changing prior to leaving a word. Additionally, we have added an option for users to force RtL only ordering to avoid bidi behavior when desired: http://arabetics.com/java/NAIM.html

Further discussion of this method and other related topics are included in our latest article in the upcoming issue (40.3) of Visible Language: “Typography behind the Arabetic Calligraphy Veil”.

Thanks.

-Saad

hrant's picture

Na'eeeem. :-)

hhp

AzizMostafa's picture

US Utility Patent Pending?! Congratulations!

Does NAIM virtual keyboard dispaly Initial or Standalone Arabic Glyphs?!
Are you going to market NAIM physical keyboards along with NAIM Software?!
Only then, New learners would not quit the educational process early due to the 2 ‘confusing’ shapes needed to be memorized up front.

Then, Arabic is a Digit Hopping System. Arabs read and write 2-digit numbers from right to left if the right digit is not zero, whereas 3-digit numbers are written from left to right but hop read from the left significant, right significant then middle...
So how does NAIM solve the insolvable?

Congratulations again with Flowers

Saad Abulhab's picture

Thanks Aziz

>>Does NAIM virtual keyboard display Initial or Standalone Arabic Glyphs?!

Both, it is not dependent on that. I made sure to test it with 2-4 glyphs per letter too. It works with any font.

>>Are you going to market NAIM physical keyboards along with NAIM Software?!

For now we created a java applet that serves fonts, keyboards, directionality, and input methods, over the web to demonstrate our project. But NAIM is just a method. It can be implemented within hardware or software, or even using OT features (with application software aware environment)

It's main goal to make a user and learner not overwhelmed by awkward glyph changing so he/she can concentrate on learning the shapes.

>>Then, Arabic is a Digit Hopping System. Arabs read and write 2-digit numbers from right to left if the right digit is not zero, whereas 3-digit numbers are written from left to right but hop read from the left significant, right significant then middle…
So how does NAIM solve the insolvable?

Funny you say that Aziz. That is exactly what we have been saying for years: bidi did not solve the insolvable? It just made typing Arabic awkward, and added unneeded overhead.

-Saad

AzizMostafa's picture

Well Saad, with the current M$ windows environment, the new learner sees on the screen what he/she first hits on the keyboard, A stand-alone Letter on both?!
But with NAIM he/she does not see what he expects by the first hit. Instead he/she will see a different shape — the initial form of the letter— unless of course he/she spaces it for example? Right? Or have I made more funny questions?!

Numberwise, I do not see any awkwardness in the current Arabic System. Numbers go naturally and smoothly as long as you are in the right (to left) direction that has been made not to toggle by accident as before. Mistakes in numbers can be easily corrected as well.

Thanks with Flowers

Saad Abulhab's picture

I see your point.

But as a comparison, one sees "Q" but when typing sees "q" on the latin keyboard. I do not think this is too confusing.

>>Number wise, I do not see any awkwardness in the current Arabic System.

I have witnessed many confused Arabic numerals users. They type "78" but see "87". As you and I observed, it is an unsolvable problem because Arabic numbers are typed 50% RtL and LtR. So why complicate our life. Let’s type numbers like we did in the Arabic typewriters. But lets have bidi as an option *not default*

I think bidi was injected "as requirement" for mixing Latin with Arabic not for the numbers.

-Saad

Si_Daniels's picture

>But as a comparison, one sees “Q” but when typing sees “q” on the latin keyboard. I do not think this is too confusing.

On most full size keyboards the UC letters are positioned at the top left part of the key designating they are accessed via the 'shift' state.

Cheers, Si

Saad Abulhab's picture

>>On most full size keyboards the UC letters are positioned at the top left part of the key designating they are accessed via the ‘shift’ state.

Today, keyboards in the US show UC only. I ordered thousands since the eighties. Typewriters did show both as I remember.

Still I am not objecting to using keyboards with non isolated shapes. In our virtual keyboards, http://arabetics.com/java/NAIM.html , we chose (to save time and for speed) not to use gifs for keys. Instead we are generating them via XML using system fonts. We could force non isolated shapes if font embedding on the web was decent. As you can see we had to choose Java to embed our fonts.

-saad

Si_Daniels's picture

>Today, keyboards in the US show UC only.

Probably beacause most 4+ year olds know the connection between the U & lc shapes of the letters - why add noise to the keyboard by adding the lc? However the connection between '*' and '8' is not so obvious hence the regular and the shift state both being printed ont he keys.

>keyboards in the US

Which country shows the lc?

Saad Abulhab's picture

Si wrote:

>>Probably beacause most 4+ year olds know the connection between the U & lc shapes of the letters - why add noise to the keyboard by adding the lc? However the connection between ‘*’ and ‘8’ is not so obvious hence the regular and the shift state both being printed ont he keys.

I think you have missed the point, Si. I replied to Aziz's post which stated that showing an isolated form Arabic letter on my virtual keyboards while displaying a different initial shape, was confusing. I said:

"But as a comparison, one sees “Q” but when typing sees “q” on the Latin keyboard. I do not think this is too confusing"

I am not proposing the addition of LC to current US keyboards. We are basically in agreement.

-Saad

AzizMostafa's picture

At 5:40am, Saad Abulhab said:
It’s MAIN GOAL to make a USER and LEARNER not overwhelmed by awkward glyph changing so he/she can concentrate on learning the shapes.

3+ hours later, Saad Abulhab said:
But as a comparison, one sees “Q” but when typing sees “q” on the latin keyboard. I do NOT think this is TOO CONFUSING.
_______________________
Now I come to understand how eye-opening NAIM is!
A must-to-have not only by the 4+year old but for the 40+year also.
US-Utility-Patent-Pending.

Flowers to All

Si_Daniels's picture

> I am not proposing the addition of LC to current US keyboards. We are basically in agreement.

Wasn't arguing just offering my explanation for Latin keyboard oddness. Still interested if any language prints lc (or equiv thereof) decals on their key caps.

AzizMostafa's picture

Sii, Arabic Keyboard shows English UC on the top left and Arabic Standalone letters on the bottom right part of the key.

So, by adding noise, I guess Saad means adding the initial forms of Arabic letters on the top right part of the key, as in old English-free Arabic Typewriters where most letters were represented by 2 forms save for 3'ain (and Ghain) 2 keys were used: one key for the initial and isolated final form and another key for the middle and connected final form.

Saad Abulhab's picture

Aziz wrote:

>>Now I come to understand how eye-opening NAIM is!
A must-to-have not only by the 4+year old but for the 40+year also.

NAIM is an input METHOD and OPTION. OPTION means it is not a MUST. Except for those who are unable to see more than one color. Our JAVA utility is a *prototype* demonstrating that method not proposing a new keyboard. It was you who complained that on our VK we had isolated forms and started the discussion about keyboards not the method. We explain to Si that we had to use the default commonly used Arabic keyboard shapes because XML was the least painful way to generate the keys via web. Our comments about "q" and "Q" do not contradict each other, they were in response to your complain.

>>US-Utility-Patent-Pending.

Let us add to your information that U.S. Patent 4,670,842 awarded years ago to Metwaly of IBM was for creating a method that does what NAIM does differently. If Giant IBM thinks it is worth patenting, why should I be bothered by your sarcastic noise? (if you would excuse our straight forwardness -:)

-Saad

Si_Daniels's picture

>If Giant IBM thinks it is worth patenting

IBM holds more than 40,000 patents (see) and generates a billion dollars a year licensing them - I'm sure they'd patent their own grandmother if they thought they'd get away with it. :-)

hrant's picture

The way the patent office works (using the term very loosely indeed) they could get away with it. It's just that granny simply isn't a marketable product. Anymore.

hhp

AzizMostafa's picture

Saad, I loved your straightforwardness since our first communication to this point of time.
How happy-go-lucky you are with the XT2XP-Screw and Throw G.I.A.N.T IBM.
Peace with Flowers

Miguel Sousa's picture

Saad, I tried the webpage with the latest Firefox (v1.5.0.6) on Mac, and it doesn't work. Is there any change you can fix this? Thanks.

Saad Abulhab's picture

>>I’m sure they’d patent their own grandmother if they thought they’d get away with it. :-)

Isn't that the same with the rest of them? Just take a look at the Adobe products, 100s of patents for each product -:)

>>Saad, I tried the webpage with the latest Firefox (v1.5.0.6) on Mac, and it doesn’t work. Is there any change you can fix this? Thanks.

Miguel, I am working as we speak on the Firefox on the Mac after wasting my time on Safari. This Java/swing project, which I have worked on since March, serves antialiased fonts to display and edit with, keyboards & keymaps, input and ordering methods, email text, and more. It was a challenge to make it work on two browsers on Windows!
http://arabetica.com/applet_info.html

-Saad

hrant's picture

Saad, IIRC, Bitstream has/had something called Jet, which was Java code for typography. I actually licensed it once (it was free, at least at launch) but never got around to using it, but I remember it was Unicode and pretty savvy (including bidi I think).

hhp

Saad Abulhab's picture

I remember it years ago. I was hopeful to use it but I gave up because it was taking forever. The first similar applet was done by Xerox, then another by IBM, then an Israeli company, Slangsoft, out of business since 2000?, also had such applet.

But in my project I need more than bidi which is the default java behavior. We needed to undo the *anti option* limitations imposed on Arabic by current solutions. Our applet does plenty more.

-Saad

AzizMostafa's picture

Saad, with NAIM, I sincerely recommend marketing 2 kinds of Transparent Stickers: One kind with 2 letters per key for the English-only PK and another with initial letters for the English-Arabic PK.
Hope I wont be accused of adding "Sarcastic noise" again.
Happy Marketing

Saad Abulhab's picture

Salam Aziz. Eid Mubarak

I have noticed that you are Iraqi too, like me. You attended Univ of Basrah 76-79. I attended Univ of Baghdad 76-79 too. Another rare 58 survivor who quits in that pre disaster year, 1979!

Thanks for the advice. I work few hours a day for free just to serve the great Arabic script. Marketing? we can not afford making business of typography. That requires money and full time attention.

In sha'a aalah, in a next update I will make two version of the virtual keyboards.

-Saad

AzizMostafa's picture

May I extend the same sentiments to you also.
Apologies with a Postcard

Syndicate content Syndicate content