Help to set a sequence of keystrokes for Ẹ̀, ẹ̀, Ọ̀, ọ̀, Ẹ, ẹ, Ọ, ọ.

[Luis de la Orden]

Hi all,

I am new to Ukelele and would really appreciate your kind help with this please.

The question is also relevant for letters that can receive two diacritics (Ẹ̀) either by themselves or both at a time.

I am trying to extend a UK keyboard to be able to type Ẹ̀, ẹ̀, Ọ̀, ọ̀, Ẹ, ẹ, Ọ, ọ in a natural, intuitive and easy way for anyone without any knowledge of keyboard layouts (the layout will be pre-installed). 

I did some research with real users and discovered that very few people will use Alt and just for very few characters, as long as it is obvious to be a clear alternative of the key they are typing and they have been taught. Combinations that require the use of two or more within the group of Alt, Shift beyond the capitalisation function or CTRL/CMD are very difficult and counter-intuitive for people who are not using the keyboard for specialised tasks (e.g.: programming, linguistics, etc). 

We are talking about mom and pop level of use here. From the easiest to the most difficult:

1. Very easy 

1.1. [( ` ) + ( o )] = ò, because the characters and diacritic are in the keyboard already and the combination produces a character that matches the keys and order typed: "ò".

2. Easy (under conditions of training and familiarity)

2.1. [( Alt ) + ( o )] = ọ, the conditions to make this use to use is that the user needs to be told about the key combination beforehand.

2.2. [(Shift) + ( Alt ) + ( o )] = Ọ, as above this will require users to realise that the shift key capitalisation behaviour continues as they know it and doesn't change despite the addition of the Alt key. They have been told that the alt key invokes their "secret" but familiar "ọ".

Ideally if a UK keyboard had a diacritic "." things would be much easier (in part as we still need to add the "`") but as far as I can see, changing the stop mark behaviour would be very counter-intuitive for users.

3. My question

3.1. How do I combine actions 2.1 and 2.2 above in Ukelele? What is the Dead key programming procedure?

3.2. Is it possible to create a keyboard shortcut or dead key in which the character combination [( ` ) + ( . ) + ( o )] = ọ̀? (this would be super easy in user terms)

3.3. Anybody know a way of making a bundle and installable DMG file with the ability of setting up keyboard character combinations automatically?

Many thanks!

Luis Morais

[Sorin Paliga]

Hello

Perhaps you may have a look at my keylayout US Academic, free download here:

http://www.unibuc.ro/e/prof/paliga_v_s/soft-reso/

That keylayout allows to either get precombined chars (e.g. é, ů, å, ł), i.e. those codified as such by unicode.org; or to combine any  letter with one, two or even more diacritical marks, those encoded by unicode.org as Combined Diacritical Marks (CDM). This latter situation is useful for all those who need linguistic or dialectal transcriptions. In order to achieve this, unlike the precombined chars, you need (1) a font containing CDMs and (2) an application able to deal with CDMs–both conditions must be met simultaneously. Fortunately, more and more apps can handle CDMs now: Nisus, Mellel, the new version of Pages (5.x, not not the older versions), iCalamus (but not other DTP apps); note that LibreOffice, OpenOffice and MS Office cannot handle CDMs (just like the old generation of Pages).

In my keylayout, CDMs are at the level option-D (CDM below) and option-shift-D (CDM above). For example, if you need to get letter O with dot below and also marking its long or short value, i.e. (using Mail.app, also supporting CDMs)

ọ̄̆

you will need to type o, then option-D followed by dot [point] then option-shift-D followed by a [long vowel], then again option-shift-D followed by b [breve]. 

It may seem difficult, but this is an issue of getting accustomed to a series of combinations. Named keylayout practically covers all the diacritical marks available, and you may get any combination you wish. As no one needs all the possible chars, practice that keylayout with the keyboard viewer, and proceed. I g

On Nov 10, 2014, at 1:32 PM, Luis de la Orden wrote:

Perhaps it is easier to start with an existing keylayout, mine, rather than building one from the scratch. If you have other needs, you may modify what you think best fits your needs.

--
You received this message because you are subscribed to the Google Groups "Ukelele Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ukelele-user...@googlegroups.com.
To post to this group, send email to ukelel...@googlegroups.com.
Visit this group at http://groups.google.com/group/ukelele-users.
For more options, visit https://groups.google.com/d/optout.

[Tom Gewecke]

On Nov 10, 2014, at 4:32 AM, Luis de la Orden wrote:

> I am trying to extend a UK keyboard to be able to type Ẹ̀, ẹ̀, Ọ̀, ọ̀, Ẹ, ẹ, Ọ, ọ in a natural, intuitive and easy way for anyone without any knowledge of keyboard layouts (the layout will be pre-installed).

What language are you actually talking about?

[Luis de la Orden]

These are Yoruba characters that I want to use in a UK keyboard, Tom. 

Just started looking at combining characters again which under the U.S. Combining layout work just as I would like, except that I want to make it straightforward for popular use. The instructions provided by Ukelele were not clear with regards to the code used so my first trial yielded no results whatsoever.

Professor Paliga, I would be very happy to see your layout, please forward a link, as the one provided is broken. Many thanks!

[Tom Gewecke]

On Nov 10, 2014, at 6:20 AM, Luis de la Orden wrote:

> These are Yoruba characters that I want to use in a UK keyboard, Tom.
>

I did an experimental yoruba keyboard a few years ago which may be helpful in your project. You can get it at

https://dl.dropboxusercontent.com/u/46870715/k/Yoruba.keylayoute

[Luis de la Orden]

Thanks Tom, but unfortunately the link is broken.

[Sorin Paliga]

It is, but if you remove the final e in keylayoute, it works!

Then copy-paste the file into TextEdit, and save it as text, with extension .keylayout

[Tom Gewecke]

On Nov 10, 2014, at 6:56 AM, Luis de la Orden wrote:

> Thanks Tom, but unfortunately the link is broken.
>

Sorry, don't know how that extra e got into the url which i copy/pasted:

https://dl.dropboxusercontent.com/u/46870715/k/Yoruba.keylayout

[Sorin Paliga]

The links to my keyboard layouts have been recovered. On this occasion I have also updated some readme files and cleaned up some stuff.

[Andrew Cunningham]

As you iindicate option and altgr key sequences tend to be anti-intuitive for end users with limited exposure to technology, it is even more problematic on the windows platform where aplications can clobber and override certain key sequences.

staying within two levels/layers is optimal.

for African language users with limited exposure to technology, we found dead keys where also anti-intuitive. people expected to type like they write, and they dont write the diacritcs before the base letter.

some of the more interesting and easy to use early vietnamese input systems actually allowed you to type vowel and tone marking at the end of the word rather than just after base vowel

deadkeys alsotend to double load the function of a key. Ie the same key can be used for either a punctuation mark or a diacritic.

how the punctuation mark is produced is in consistent and varies across keyboards. it may be produced typing key twice, or by typing key then the spacebar , or the key and some neutral key not otherwise used as a base character in a deadkey sequence.

with diacritic heavy languages, esp when mixing combining characters from classes 220 and 230 (below and above base character) though needs to be given to wether consistency is needed, whether out put needs to be normailised in some way, wether the input sequence options will be constrained to a specific order, whether input needs to play nicely with spellchecking, etc.

A.

--
You received this message because you are subscribed to the Google Groups "Ukelele Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ukelele-user...@googlegroups.com.
To post to this group, send email to ukelel...@googlegroups.com.
Visit this group at http://groups.google.com/group/ukelele-users.
For more options, visit https://groups.google.com/d/optout.

--

Andrew Cunningham
Project Manager, Research and Development
(Social and Digital Inclusion)
Public Libraries and Community Engagement
State Library of Victoria
328 Swanston Street
Melbourne VIC 3000
Australia

Ph: +61-3-8664-7430
Mobile: 0459 806 589
Email: acunn...@slv.vic.gov.au
          lang.s...@gmail.com

http://www.openroad.net.au/
http://www.mylanguage.gov.au/
http://www.slv.vic.gov.au/

[Luis de la Orden]

Hi fellas,

Good point Andrew! The information on typing the letter before the diacritic is very valuable as we need to take into consideration the way people would type and how intuitive and learnable a combination of keys is.

I have been busy working on a first draft for a UK MacBook Pro (2013), not there at all but just showing how all your comments and examples have been helpful to get me started!

https://www.dropbox.com/sh/swee53872npxkoq/AAB4HfAJFJFEZK_gOgUNgmvba?dl=0

Many thanks to Professor Paliga and Tom, if I haven't said so already. I believe that if we carry on with this discussion we can come up with something really useful for people willing to write (in Yorùbá at least) from UK, US and US international keyboards.

I am talking to other people with regards to the best combination of diacritics and dead keys to produce characters such as ẹ̀ and ṣ without allowing for aberrations such as f̣̣̀̀, as Patrick Chew pointed out somewhere else. Please keep sharing your opinions.

Cheers,

luis

[Luis de la Orden]

Hi fellas,

With regards to the keyboard layout available in my link above, you can access the diacritics by using the option key, sorry it is still that basic. As Yoruba uses the same latin letters found in our keyboards, and most Nigerian students and writers use either the US international or UK layouts it is the accents already found on the keyboard (initially the British one) that I want to reconfigure to allow `, ' and . to produce characters such as ẹ́ and ẹ̀.

Apologies for any confusion,

Luis 

[Sorin Paliga]

Let us clarify some principles, perhaps useful.

The norm is to use option/alt key for getting diacritical marks others than the ones common in a given language. English has no diacriticals, other languages have a limited number (Finnish, Swedish, German), others a lot (Czech, Polish, Lithuanian etc.). The principle is to have as many as possible precomposed chars (letter + diacritical) at the lowest level possible, because it would be particularly uncomfortable to press option key all the time when writing, say, in Czech or Polish. This is the basic principle when currently using a language, for which the norm specifies diacritical marks, disregarding their number: they must be available at the lowest level possible.

Things are different when dealing with linguistic or dialectal texts, in which we use or any other language, and such ‘eerie’ words must be inserted from time to time. In such a case, it is of course more comfortable to have a basic keylayout allowing a rich list of letters + diacritical marks. This was the philosophy of my US Academic or other similar keylayouts. The list of possible combinations is huge, but no linguist/dialectologist needs all of them, but a limited number. In order to learn them, he/she must practice several minutes a keylayout with Keyboard Viewer active, until he/she learns the combinations. It may seem tedious or slow, in fact it is not. This method is, in fact, a lot faster than switching keyboards. Of course, this is NOT recommended when writing in Czech or Polish or Yoruba, for a current use there must be a keylayout designed for each of these languages. 

It is yet practical to have, say, a keylayout for Czech if writing long texts in Czech, but this keylayout should allow an easy access to other non-specific chars, e.g. inserting Serbian or Polish or Romanian or Lithuanian names in a Czech text, the keylayout for Czech should allow writing these chars via option keys, as these words, usually personal or place-names, are rare, they occur from time to time. That was the philosophy of my keylayouts for Romanian or Czech, while Cyrillic Linguist allows to write Cyrillic words with various chars specific to non-Slavic languages of the former Soviet Union, Uralic or Altaic languages for which Cyrillic has been the norm.

[sil.linguist]

Luis and Andrew C.

You both mention testing users. I am interested in knowing more about your methods. I too am working with African orthographies with diacritics. @Andrew I think I have seen some of your materials online and read them.  I am also interested in finding out more about how this Word-Final marking of tone in Vietnamese works. I have wondered if there is a text input system like this.

@Professor Paliga I am wondering if you have any academic citations, or ISO/Standards body citations for the "principles" you outline. They seem rather straightforward but certainly they are documented in some text somewhere?

If it is more convenient to contact me via PM please do so. If it is even more convenient to skype or use google hangouts please, lets set something up.

- Hugh Paterson  - 

email is FirstName [dot] LastName [at] SIL [dot] org

[Sorin Paliga]

On 13 Nov 2014, at 08:40, sil.linguist <sil.li...@gmail.com> wrote:

Luis and Andrew C.

You both mention testing users. I am interested in knowing more about your methods. I too am working with African orthographies with diacritics. @Andrew I think I have seen some of your materials online and read them.  I am also interested in finding out more about how this Word-Final marking of tone in Vietnamese works. I have wondered if there is a text input system like this.

@Professor Paliga I am wondering if you have any academic citations, or ISO/Standards body citations for the "principles" you outline. They seem rather straightforward but certainly they are documented in some text somewhere?

Standards, of any kind, refer to technical issues, e.g. what specific chars with specific encodings should be present in a keylayout, they NEVER specify to press second or third finger to get a char. 

I have the feeling of a kind of aggressive phrasing, accusing me of lack of citations where no citation should be. As an example, could you cite anything specifying why ß is got when pressing option-S and Í if pressing option-shift-S? (U.S. basic Apple keylayout). 

To use mnemotechnical means and have simple solutions is not specified in ISO documentation, but if you wish to write something on this issue, please do, I will surely quote it.

[Luis de la Orden]

Hi Sil.Linguist, Prof. Paliga, Andrew,

 

I will try to summarise the normal procedures I use in Human-Centred Design, which can be transferred to other areas of design (such as keyboard layouts, etc…). I use these practices in software and systems design and in fact they are all borrowed from engineering, psychology, architecture and so forth.

 

1. Identification of the target audiences their needs, habits and hardware

 

It is very obvious that everyone in this forum deals with this step by knowing who will use their keyboard layouts, level of technical knowledge, for what tasks and the base layouts that will be changed by the new layout.

For example mine are:

a. Nigerians that use English as their main writing system, living in Nigeria, the US or the UK who want to write in Yoruba either full time or in the middle of English text. UK/US keyboard. Average technical knowledge. Need to have it installed for them; 

b. Brazilians who own a US or UK keyboard, are interested in Yoruba culture or are descended from Yoruba stock and want to use Yoruba accents and tonals to understand religious texts an songs. Good technical knowledge. Can follow instructions to install;

Both audiences a and b come from very distinct contexts. The context of Brazilian use is one of the individual who actually can understand English and afford to make a choice of buying a US MacBook Pro (it is cheaper than buying it in Brazil itself, another social factor to know) and doesn’t feel uncomfortable of using a foreign layout. They will use it for religious purposes, since we have a strong following of the Yoruba religion in my country combined with a recent push to re-learn that tradition properly, starting with the diacritics, which help us pronounce the words correctly. So other needs to fulfil are religious and pedagogical.

Needless to say that both contexts are more Windows users than Mac users, but we already established that will be followed up later.

Andrew shared an interesting piece of information when he mentioned that Africans with *little technical knowledge* would not expect to press an accent and then write their letter because they are accustomed to first writing the letter and then the accents.

This statement wouldn’t work for Brazilians, as we are already used to first pressing the accent and then the letter to produce the fused accented letter. This is what we have been taught by typing machines of yore. So ethnography counts in identifying whether previous technology had already established certain habits. The more people you want to use your layout the more you need to know about them. 

 

2. Standards, best practices and conventions

Prof. Paliga’s explanation on what should be at the very first level of typing is a good example of a best practice or principle of design.

Best practices are backed by

Human factors:

a. e.g., we have 10 digits (sometimes less),

b. digits don’t bend sideways,

c. one digit is easier to use than two, two digits are easier to use at the same time than three;

 

psychological and cognitive factors:

a. E.g., Group functional keys instead of mixing them randomly;

b. Make combinations intuitive and consistent (‘ and vowel will always produce a consistent character output);

c. Avoid new or rare paradigms (the use of the option key is unusual, but sometimes it is required;

d. Where new paradigms are introduced make them learnable and consistent (it is all about consistence);

e. etc, etc,

 

technological/technical factors

a. e.g., what hardware will be used;

b. what other alternatives are available; etc.

c. what is the mean age of software used by most people; can it support Unicode;

In my practice with software usability, most of this knowledge you accumulate by using and testing with/observing people using the system.

ISO is good to give one the benchmark to assess how effective a system is to address the human factors and cognitive factors under several technical contexts. What they do is to pick the best practices and stuff that is proven to work and turn them into high-level standards to be followed as principles of application, they will not give specifics on how to achieve compliance, this is learnt somewhere else with industry practices manuals but ISO is still useful with regards to giving the product creator an idea of what they need to achieve in order to have a product that is suitable for a wide range of human needs and applications.

I think it is a great idea to have something practical written about the subject of keyboards. Jakob Nielsen is the most notorious usability engineer people in my field grew up reading to: http://www.nngroup.com/articles/movies-usability-top-10-bloopers/.

 

3. Prototype and test with users

 

Given that we have identified our audience.  

When you test you will check for these to be in order:

a. Affordance and intuitiveness

Can people tell straight away or in a few trials without being given any clues except the information that the keyboard can produce those accents where they can achieve the accented output? Can they guess?

 

b. Learnability

Failing the above, not all is lost. Can people learn a formula of use easily and apply it over and over to obtain the results they expect? Can they remember it easily? Can they teach other people?

c. Recoverability

Can they correct their own mistakes? Do they keep making the same mistakes even after training?

The way to test is not that difficult, you put the keyboard they are familiar with in front of them and interview them starting from the way they type their most used language in that keyboard and progressively ask them to try to discover the key combinations taking notes of what they did by themselves. 

Proceed to teach them the combination formulas and ask them to apply the principles themselves, take notes. Teach a bit more ask them to do it by themselves. Rinse, wash, do it again.

Ask them to type a sample text, one with the language and key combinations they normally do, another one with the new shortcuts and finally a third with both the familiar and new. Ask them for there opinions and make sure they are not trying to please you. Be Freudian and do not put words in people's mouths and in 2 - 4 iterations you should have an idea of how efective your layout is for the audience intended.

I hope this may help,

Luis

[Hugh Paterson III]

Luis, 

That is really helpful. It is good to know that you are taking a UX approach here. I too am in the UX field.

This is a great outline. I am still curious about what metrics are used to evaluate mis-types, or learnability difficulties, typing speed, etc. Are you (and Andrew) using key logging or just observation? perhaps post use interviews? Eye tracking or video of typing?  Rather than the highly educated target audience you are focusing on, I have, like Andrew the quality in my target audience *little technical knowledge*. 

The contrast in expectation between your audience and Andrew’s audience is an interesting one, and likely deserves some more investigation, comparison and reporting - even to pose the question, does learning another language in a digital context mean learning new ways of text input?

BTW: I recently published a chapter on Keyboard layout design here: http://bit.ly/1sIGJim

- Hugh

--
You received this message because you are subscribed to a topic in the Google Groups "Ukelele Users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ukelele-users/G2-J612Su7M/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ukelele-user...@googlegroups.com.

[Sorin Paliga]

Beyond any doubt, yes: a keylayout must have a target group of users. Those included in the system cover most needs, even if they have (and some do have) flaws or missing chars.

A completely different category is represented by software developers, who prefer Programmers-type keylayouts (it is clear why).

And another category of users is represented by linguists, dialectologists and any in this category, they often need a large number of chars to be inserted in a linguistic analysis. A subcategory is represented by those who need a frequent use of some non-standard chars at zero and first level (no modifier key pressed or shift key pressed), they need a custom keylayout, e.g. the one for Yoruba or any other language, which makes use of a large number of chars.

We cannot invoke ISOs and other norms here, it is a question of ease of use, encodings are already clarified and present in the unicode.org lists.

You received this message because you are subscribed to the Google Groups "Ukelele Users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to ukelele-user...@googlegroups.com.

[Luis de la Orden]

Hi Hugh/Sil.Linguist

This is a great outline. I am still curious about what metrics are used to evaluate mis-types, or learnability difficulties, typing speed, etc. Are you (and Andrew) using key logging or just observation? perhaps post use interviews? Eye tracking or video of typing?  Rather than the highly educated target audience you are focusing on, I have, like Andrew the quality in my target audience *little technical knowledge*. 
The contrast in expectation between your audience and Andrew’s audience is an interesting one, and likely deserves some more investigation, comparison and reporting - even to pose the question, does learning another language in a digital context mean learning new ways of text input?

Apologies for the delay in replying. 

Any form of recording such as keylogging and recording are very useful. Keylogging is much more useful with video recording of the user action, mostly with tasks such as typing as you want to visualise how fingers "dance" and "flow" in your keyboard layout in real world terms. 

Ok, very likely you will not have any previous data to create your benchmark so that the level of effectivenes and learnability can be analysed (for sure there are loads of information on the matter hidden in the secret vaults of corporations which will not divulge anything as a matter of trade secrets and competitive advantage). Nevertheless you can still create your own benchmark and start from there, this is pretty much what everyone does.

The formula is to always document what peole are already doing and analysing that ethnographically. In the case of Yoruba, I discovered that there is a lot of cut and paste going on and most people think that this method suffices them without realising that they are much slower and less effective writers. At this point you don't want to convince anyone but also be aware that people won't know it better, so don't ask for solutions fromt he people interviewed but record those rare conscious complaints and the much more rich between the line behaviour when you ask them to do the tasks int he way they are used to. By the way, the very first few 4 or 5 people you start to test will not make to the statistical analysis as you will discover  that these first few people will give you ideas to better your own study methodology, by the fith person you will have added or taken things you originally designed as part of the test.

Yes, the same principle applies for novel input methods but if they are against human factors and usability, it will be a waste of time to insist with testing people as the answer is already in biology. As an example, we know of a company (https://www.leapmotion.com/) whose product offering is based on the muscle-tiring and inaccurate handling of the screen via hand waving and poking in the air. Minority Report. I used the bloody thing and in 15 minutes my arms were killing me and I couldn't do anyhting useful such as typing or selecting text. They are selling loads but I doubt anyone is using their LeapMotion more than a few days and then boxing the thing again or passing ahead to friends, in the hope they will find it useful.