Final Report on AYNISIWI

deliriousattackInternet και Εφαρμογές Web

4 Δεκ 2013 (πριν από 3 χρόνια και 11 μήνες)

97 εμφανίσεις


1

Final Report on AYNISIWI


Iván Guzmán de Rojas

UNESCO Fee Contract No. 4500014108

May 2004






1.

Introduction




2.

Usage and data collected




3.

Analysis of the effectiveness of the translation tool




4.

Analysis of the process




5.

Lessons learned




5.1
Operational issues




5.2
Technical issues




5.3
User interface






6

Recommendations





6.1
Enhancements to AYNISIWI




6.2
Continued Operation and Support





Annex 1
Participants registered at Aynisiwi_b@bel (2004/05/24)




Annex 2 Posting
s in English



Annex 3 List of missing terms (Source text: English)



Annex 4 Postings in Spanish



Annex 5 User feedback
-

Form







2


1.

Introduction

Within the framework of the
b@bel

initiative aime
d at promoting linguistic
diversity and equitable information access in the digital environment, Iván
Guzmán de Rojas with his IGRAL R&D group in Language Engineering, have run
an experimental operation of the AYNISIWI multilingual online forum to "field
t
est" the translation engine ATAMIRI, for the 5 languages: English, Spanish,
French, Portuguese and German in all 20 translation directions.


The objective of this experiment was a fully operational demonstration of the
feasibility of linking a web
-
based on
line forum tool to the ATAMIRI automatic
translation engine thus enabling the exchange of information and view points
between users working in several different languages. This combination of
software permits the interaction of people who are separated bot
h in space (i.e.
geographically) and in time since the communication is asynchronous (in contrast
to “chat” software that requires users to be online simultaneously).


The operation was run from 15
th

January 2004 to the end of May 2004. It took
place at t
he web site:
www.atamiri.cc/aynisiwi_b@bel


The forum has been used to support the development of multilingual terminology
and specialized dictionaries by “mining” forum postings and evaluating the
reliab
ility of the translation engine and user satisfaction of the outputs.


3

2.

Usage and data collected

36 participants, 15 contributed with at least one posting (see annex 1).


Country


Number of participants

Bolivia



11

France





6

Canada





5

Spain





4

UK




2

Suisse





2

Guatemala



1

Paraguay



1

Uruguay



1

Sweden



1

US




1

Ethiopia



1


Language


Number of postings (31/05/04)

Number of words



F1

F2

F3

F4

All


English


23

16

9

13

61


7,212

Spanish


4


7

2


2

15



1,035

Fr
ench



1


0

0


0


3



199

Portuguese


0


0

1


0


1



18

German


1


0

1


0


2



74


The postings in English are shown in annex 2, and those in Spanish in annex 4.
The translated postings from English into Spanish are shown in annex 2a
,
including in brackets [ ] the assigned intelligibility factor.

The lexicographic analysis of the postings in English has allowed us to do the
“mining” of missing terms which are shown in the list in annex 3, with the
following statistics:


Language

Numbe
r of missing terms

% of missing terms in text

English




80




1.04

Spanish



85




1.17

French



145




2.01

Portuguese


372




5.16

German


182




2.52


4

3.

Analysis of the effectiveness of the translation tool


Before introducing the missing terms in the
lexicographical data base, we will
analyze translation effectiveness in all 20 given directions using the following
approximate method:


If the translated posting was intelligible it got 10 points. If the posting was not
quiet intelligible it got a number

of points between 5 and 9. If the posting was
unintelligible it got a number of points between 0 and 5. For each translation
direction, an “intelligibility factor” has been calculated as the average number of
points obtained by the corresponding number
of postings translated in that given
translation direction. We started first with both translation directions EN <
-
> ES
and obtained the following preliminary results:

Translation direction




Intelligibility factor_________________________




Forum 1


Fo
rum 2


Forum 3


Forum 4


All

EN
-
> ES


8.0


7.7


6.1


8.7


7.6

EN
-
> FR




EN
-
> PT




EN
-
> DE





ES
-
> EN


8.3


7.0


6.5


7.5


7.3


ES
-
> FR




ES
-
> PT




ES
-
> DE





FR
-
> EN


7.0


--


6.5


--


6.8

FR
-
> ES


9.0


--


7.0


--


8.0

FR
-
> PT




FR
-
> DE





PT
-
> EN




PT
-
> ES




PT
-
> FR




PT
-
> DE




DE
-
> EN




DE
-
> ES




DE
-
> FR




DE
-
> PT




Key: EN English, ES Spanish, FR French, DE German, PT Portuguese


5


4.

Analysis of the process

The AYNISIWI forum software is written in Java and manages th
e interaction
with the user through common web browsers (such as Microsoft Explorer) using
a combination of standard HTML, forms and javascript. AYNISIWI stores the
original text and language of messages in a tree structure organized by topics
and replies
to previous messages. When displaying the contents of a message to
a user, if the user is working in a language different from the original text, the
original text is passed to the ATAMIRI server (which can be located on a remote
server different from AYNI
SIWI itself) for translation. The results of the translation
are returned to AYNISIWI to be displayed to the user alone with the original.

Lexicographic analysis of the results of translations can be used to guide the
update of the ATAMIRI translation tabl
es through another web
-
based user
interface (ARUNQERA). The results of these enhancements are immediately
available to AYNISIWI users as improved translation quality.


Login screen


Participants were recruited through a letter of invitation sent to people

known to
have an interest in automatic translation and in most cases having some
familiarity with more than one of the languages currently supported by ATAMIRI.
Participation was openly and informally extended to others by the individual
initiative of the

initial users.


6


Discussion was initiated by the moderator (Iván Guzmán de Rojas). High level
forums were defined (see screen view below). Then initial explanatory messages
were posted under each topic.




High level forums



Users were asked to post mess
ages according to their personal interests and in
the language of their choice. Other users could read and reply to messages in
the same manner as other web
-
based forums, except they were free to do so in
their own language, independent of the original lan
guage of the message.


In the forum
Help & FAQs

the functionality of the system was explained, some
operational instructions were given, and the required assistance related to help
questions was also provided




7


Topics in first forum


Topic titles were or
dered by date of posting (last entered first). They appeared in
the screen in the original language and if necessary with the translation to the
language of the participant declared at registration time.


In Aynisiwi, the data base of posted messages is or
ganized in a tree structure;
each forum contains topics; topics contain contributions; contributions may
contain comments, and so on. By clicking the (+) icon at the left side of a title you
can unfold the branch and display its lower level branches. Rever
sely with the (
-
)
icon, you can fold together the branch.


Participants are allowed to either post a new topic or a contribution to a given
topic. The same procedure applies to lower level postings (comments, notes to
comments, and so on).





8


Welcome me
ssage posted in English


This postings presentation format has proven to be adequate for a multilingual
web forum. The reader can choose the target language for translation. Per
default the translation into the participant registered language is given. I
f it is
the same as the source language, no translation is generated.


The posted messages in their original languages are stored in the Aynisiwi
data base. The translations are not stored, they are generated each time it is
required. It may happen that t
he translation quality of a given posting improves
after few days, due to the on
-
going lexicographic data base enrichment or
thanks to translator engine enhancements.


The postings may contain HTML code, but even though it is well handled in
the source lan
guage screen, in some cases it may become corrupted in the
target language, for example, link references. These anomalies will be fixed
for next Aynisiwi version.


The length of a posted text is restricted to approximately 300 words. For
example the text i
n the screen shot shown above has 274 words. Most
postings are much shorter.


9



Translated posting (from English into Spanish)


The translated postings are always shown in its raw form (not post
-
edited).
Those words marked with an asterisk are cases whe
re the translator engine
failed to recognize them, either because they don’t exist in the lexicographic data
base, or because the system was unable to handle them in that context
(anomalies which have to be corrected in the enhanced version).


Translation
quality depends on the language direction chosen. For translation
from English into Spanish it is pretty good, for example the translation shown in
the above screen shot is considered to have an intelligibility factor of 9 in spite of
some grammatical err
ors, because the message contents are well understood
without recurring to the source text.


In other directions, translation quality is negatively affected by the limited lexical
availability. The lexicographic data base contains almost 30,000 entries in
English and in Spanish, while for other languages it contains less than 20,000,
which makes a notable difference in translation quality.


10


During its operation AYNISIWI collects data in the form of transcripts and web
logs recording time and date of each u
ser section. These sources have been
used for respective analysis of the results of the experiment.



Lexicographic analyzer tool


The lexicographic analyzer of the system is a multilingual tool that allows to trace
in a given source text those words whic
h are missing in the lexicographic data
base in the languages used for translation. The above screen shot shows the
three words missing in the source text in Spanish, one of them occurs twice.
Moving the cursor it is possible to select the text sentence
where the missing
term occurs. In this way the task of lexical enrichment can take place using
context analysis of word occurrences.

During the experiment lexicographic enrichment was carried out by detecting
missing words in the translated texts and by i
nspection of translation results. In
this way ongoing incremental improvements to translation quality were made.


In addition, user feedback on the service was solicited using the forum itself as
well as an email sent to all participants. The request for f
eedback is included
Annex 5, it has been also posted in the forum. We have not yet received any
feedback, so that there are no results to be processed.


11

5.
Lessons learned


During the experimental operation of the system the following issues and
solutions

were identified.


5.1
Operational issues


An insufficient number of participants in some languages lead to difficulty in
evaluating some language directions such as German to Portuguese due to
limited data. As is observed in other web
-
based forums, the ra
tio of people who
write new postings to the number of people who read postings is very high (less
than 1 in 100), therefore a large number of users is required to fully test a system
such as AYNISIWI in a realistic manner. The issue of the number of users
is
compounded by the need to include a sufficient number of active users in each
language. For this reason we have recommended below that this experiment
continue with a greater effort made to solicit more users.


During the course of this experiment the l
exicographic enrichment could only be
done in English and Spanish due to limitations in budget and lack of available
linguistic resources. This demonstrates the feasibility of such enrichment but
more resources in this area are required if the operation of

AYNISIWI is to
continue. See recommendations below.


5.2
Technical issues


Parsing of embedded HTML coding has caused some problems, for example
links to other documents, where in some cases the contents of HTML tags were
modified in the translated versio
n, which should not happen.


Encoding of extended characters and diacritical marks required in most
languages is sometimes not handled correctly in the current version and can
occasionally result in poor translation quality. Testing has shown that this is
not a
problem with the ATAMIRI translation engine but rather due to limitations in the
interface between AYNISIWI and ATAMIRI. Both of these issues can be
addressed by improvements in the ANYSIWI program. See recommendations
below.


5.3
User interface


AYN
ISIWI lacks some features commonly found in other web
-
based forum
packages. This is not a drawback from the point of view of proving the feasibility
of the AYNISIWI multilingual forum concept; however it could difficulty in user
acceptance in the future. F
or example, in AYNISIWI it is not easy to identify new
postings since the last time one accessed the system (i.e. What's New). Also the
current version of AYNISIWI has no facility for searching for postings based on
subject and content. In these respects t
he current version of AYNISIWI is only a
“proof of concept”.


12


Since AYNISIWI is written in Java, it can be quite easily extended to include new
features. Alternatively, since the interface to the ATAMIRI translation engine is (in
principle) independent of
AYNISIWI, the multilingual features could be
incorporated into other freely available or commercial forum and portal software.


AYNISIWI has some prerequisites with respect to the user's computer system.
For example, the system may work slightly differentl
y when accessed using
different browsers, and in particular older versions of browsers might have
difficulty displaying accented and other special characters. Users might
experience difficulty keying of characters with diacritical marks if the computer
sys
tem is not configured for the specific language and keyboard options. Of
course this can be facilitated by the proper configuration of the computer.



6
Recommendations


From the experimental operation of AYNISIWI over the last four months we
consider impo
rtant the following recommendations.


6.1
Enhancements to AYNISIWI


The AYNISIWI
-
ATAMIRI interface program should be improved to reliably handle
extended character encoding in all cases. Further, the interface to ATAMIRI
should be generalized and fully doc
umented so that it will be possible to integrate
ATAMIRI's multilingual capabilities into other web
-
based systems such as email
and web content management systems.


Additional resources should be made available specifically for lexicographic
enrichment in

both terminology (missing words) and morphological coding (word
usage) in all of the currently supported languages. At least two additional
languages should be further implemented in ATAMIRI to double the number of
possible translations directions form th
e current 20 to 42. For example, Dutch
and Italian, those are already preliminarily implemented in the system.


In some languages, especially German, the grammatical tables also should be
extended to handle more cases used in informal writing.


To reliably

identify serious users, to avoid unwanted spam postings and to
validate email addresses, enhancements need to be made to the user
registration procedure. Upon completion of the personal information form,
AYNISIWI should send an email to the registrant wit
h a temporary randomly
generated password. This password would be used to access AYNISIWI for the
first time when the user chooses a new password.


Some simple improvements to the AYNISIWI user interface could be made so
that it is easier for users to loca
te topics and discussions of interest to them. It
should be possible for users to expand (+) more than one message/reply thread

13

at the same time.



It should also be possible to search for messages based on the subject and
message contents. The search must

be insensitive to the language of the
message and of the query. ATAMIRI makes this possible because of its use of
an Interlingua language
-
independent representation. AYNISIWI could be
extended to take advantage of this feature.



6.2
Continued Operation a
nd Support


A summary of the current development status of ATAMIRI is shown in the
following table, where each language in implementation is located according to
its levels of translation capability (TC) and lexical content (LX):


Language

Translation

Capa
bility

Number

of

entries

(31/05/2004)

TC0

TC1

TC2

TC3

TC4

TC5

TC6
-
9

Spanish






LX4



27,748

English






LX4



26,954

French





LX4




22,097

Portuguese




LX3





12,780

German



LX3






15,681

Dutch



LX3






11,473

Itali
an



LX3






14,329

Romanian



LX2






4,767

Russian *


LX2







9,814

Aymara


LX2







6,343

Swedish


LX1







2,644

Hungarian

LX0








1,034

Turk

LX0








167

Finish

LX0








20

Japanese *

LX0








68

Guaraní

LX0








6

Basque

LX0








10

Catalan

LX0








39

* Using Romanized representation



The advanced location of English, French and Spanish in this table explains the
better translation quality delivered

in Aynisiwi for this languages as compared to
the others used during the experiment. We recommend that an effort should be
made to raise these languages to the LX5/TC7 implementation level in order to

14

confirm that with the Atamiri translator engine it is
possible to obtain translations
with high grammatical quality and not only a good intelligibility factor.


We also recommend that it should not be missed the opportunity to implement a
native language like Aymara in this system using its capability to han
dle
agglutinant morph
-
syntactic structures. In the case of Aymara it has been already
started the lexicographic work, currently with 6,343 entries. Its grammar is also
well known to the system. Such an experiment would be an interesting first test
of an
implementation of a “minority” language, relating it to various European
languages.


We recommend that will continue to host AYNISIWI on one of Neotec’s servers
to permit an ongoing experimental operation. This is very convenient due to
direct accessibilit
y to the system at the server with good technical operational
support.


The UNESCO B@bel Initiative should advertise the availability of AYNISIWI
more widely through the UNESCO portal and other information services in an
effort to promote the use of AYNISI
WI and to invite greater participation in the
experiment. The forum main topics could be changed to include those which
would be more appealing for young people to promote intercultural debate.


Additional funding should be made available for “AYNISIWI pha
se 2” to make
technical improvements mentioned above and to continue experimental
operations with further enhancements of the translator engine ATAMIRI to
improve translation quality, which depends not only on further lexicographic
enrichment, but also on
more work in fine tuning the grammatical algorithms of
the system.



15

Annex 1
Participants registered at Aynisiwi_b@bel

(2004/05/24)* posted


01 igr



Iván Guzmán de Rojas


ES

Bolivia


*

02 phector


Paul Hector



EN

France


*


03 igornuk


Igor Nuk



EN

F
rance

04 Davide


Davide Storti



EN

France



05 Daniel PRADO

Daniel Prado



ES

France


*

06 2048123LP

Karola Guzmán de Rojas

EN

Bolivia

07 mikear


Miguel Angel Ramirez


ES

Bolivia


*


08 TDS


Tarcisio Della Senta


EN

Switzerland



09 Gunther


Günther

Cyranek


DE

Ethiopia

10 jbataller


Joan Manel Bataller Rosa

ES

España


*

11 Vaneeta


Vaneeta Mahtani


EN

UK


*

12 Gladys


Gladys Dávalos A.


DE

Bolivia


*

13 pbrand


Peter Brand



EN

Canada

14 spage333


Sheila Page



EN

Canada

*

15 billpage


Bill P
age



EN

Canada

*

16 Simon


Simon Robinson


EN

Canada

17 vfernandocv

Fernando Carreño


PT

Bolivia

18 hcastill


Hugo Del Castillo


ES

Guatemala

19 René


René Rocha Gallardo


ES

España


*

20 mauge


Maria Eugenia Aparicio Leaño ES

Bolvia

21 sfs



Sulma

Farfán Sossa


ES

Bolivia

22 Fer



Angel Ojeda



ES

Bolvia

23 patxi


Patxi Goenaga Mendizabal

ES

España


*

24 arcangel


Gabriel Mariaca


ES

Bolivia


*

25 juanal


Juan Alberto Flores


ES

Bolivia


*

26 Megavid


David Langer



EN

Paraguay

*

27 lietaer


Bernard Lietaer


EN

US


*

28 benoit


Benoit Lamey



FR

France

29 ramonfv


Ramón Fons Vilardell


ES

España

30 Bearth


Thomas Bearth


FR

Suisse

31 entel


pruebas



ES

Bolivia

32 lbarnola


Luis Barnola



ES

Canada

33 Vitelio


Vitelio Herrera



ES

Francia

34
tomh


Tom Hewitt



EN

UK

35 rob



Rob Roy



ES

Uruguay

36 titicacoki


Iván Lucksic Daza


EN

Sweden

*




16

Annex 2 Postings ordered by language

Postings in English

#1.01

Welcome to the Promoting Linguistic Diversity section of the forum!

2004
-
02
-
26 11:43:08

-

phector


According to one expert over 90% of all Internet contents exists in only 12
languages. Another expert predicts that off
-
line about 50% of the world's 6000
languages and their knowledge systems will disappear by the end of this century.

Of cou
rse the reasons for these conditions are quite complex and to solve them
one must address policy, social, economic and technical and a range of other
issues. We hope that through this forum we can discuss some of these
challenges and hopefully find solutio
ns that could contribute to overcoming the
current "language divide" and rapid disappearance of languages.

We would very much like you to use this forum to share ideas, solutions,
exchange experiences and perhaps even build a network for action. Perhaps y
ou
know of projects or studies, online resources, institutions working in this field.
Please share this with us. Are there languages not supported in the digital
environment or which perhaps no fonts are available? if so let us know, perhaps
this knowledge

may encourage someone to develop these or other language
resources or maybe someone else knows where this resource can be found.

The forum also provides an opportunity to test and refine the ATAMIRI translation
engine so your posted comments are most imp
ortant. If you wish to test the
system you can make posting on any subject in the “Field
-
test posting” section.
Do share with us your comments on the results generated by the translation
engine.

So we invite you to post your messages, share your ideas and

contribute to this
effort to promote language diversity.

#1.02

open source projects related to natural language processing

2004
-
02
-
23 21:20:57
-

billpage

The OpenNLP Homepage

OpenNLP is an organizational center for open source projects related

to natu
ral language processing. Its primary role is to encourage and

facilitate the collaboration of researchers and developers on such

projects. Click here to see the current list of OpenNLP projects. We'll also try to
keep a fairly up
-
to
-
date

list of useful
links related to NLP software in general.


For the latest news about OpenNLP and to participate in discussions,


17

check out OpenNLP's Sourceforge project page .


Notice: OpenNLP has been quite a low key "organization" (as we

envisioned it), but it noneth
eless appears to have some reasonable

recognition on the web as a place where you can get a bunch of useful

pointers to NLP resources. Not bad for putting very little time into,

but in case someone thinks they can do more to keep the web site

up
-
to
-
dat
e, I would be more than happy to pass the task on to that

person. Please email jmb@cogsci.ed.ac.uk if you are interested in fostering a
more active website!


#1.03

Language and Thought

2004
-
02
-
22 16:45:14
-

spage333


This excerpt is from an article pu
blished by Dan Slobin of the University of
California, Berkeley. The date is not confirmed but the author quotes another
source dated 1996 thus indicating an approximate time frame.

"Comparisons of different languages can lead one to pay attention to 'uni
versals'
--

the ways in which all languages are similar, and to 'particulars'
--

the ways in
which each individual language, or type of language, is special, even unique.
Linguists and other social scientists interested in universals have formulated
theori
es to describe and explain human language and human behavior in general
terms as species
-
specific capacities of human beings. However, the idea that
different languages may influence thinking in different ways has been present in
many cultures and has give
n rise to many philosophical treatises. Because it is
so difficult to pin down effects of a particular language on a particular thought
pattern, this issue remains unresolved. It comes in and out of fashion and evokes
considerable energy in efforts to supp
ort or refute it."

This has been posted as a sort of test message by me (Sheila) because it
represents to some degree the type of linguistic questions I am interested in and
I do not feel competent as yet to post such thoughts of my own.

#1.04

About Humbo
ldt

2004
-
02
-
22 17:24:13
-

spage333

Wilhelm von Humboldt, an historical figure in linguistics, created what is known
as the 'Weltanschauung' (world view) hypothesis. He believed that language
determined thought. Others have asked "How, if there was no th
ought before
language, did language arise in the first place?" Humboldt's theory may have
some linguistic validity but it certainly can not stand up to science.


18

This controversy give rise to questions such as "Which came first, the chicken or
the egg?" or

"Are able to accept some things without feeling obligated to explain
and try to prove them?"

Please discuss.

Sheila


#1.05

Some questions

2004
-
02
-
22 16:59:32
-

spage333

Help!

I do not have regular access to a computer program that allows me to us
e
anything but English. I am going to try to post a German linguistic quote but what
worries me the most is the inability to use accents.

If I post in French (or Spanish)I must also use the appropriate accents in order
for it to be correct.

What can I do

to solve this problem?

Is it possible to leave out the accents and still get a proper translation?

Advice needed, Sheila



#1.06

Some questions

2004
-
02
-
22 16:59:32
-

spage333


Help!

I do not have regular access to a computer program that allows me

to use
anything but English. I am going to try to post a German linguistic quote but what
worries me the most is the inability to use accents.

If I post in French (or Spanish)I must also use the appropriate accents in order
for it to be correct.

What ca
n I do to solve this problem?

Is it possible to leave out the accents and still get a proper translation?

Advice needed, Sheila

#1.07

RE: Some questions about accented characters

2004
-
02
-
22 23:48:35
-

billpage

Sheila,

If you are using Windows (I thi
nk the answer must be yes.) then even if you

19

computer is configured to use English as the primary language, there are still
several ways to enter accented characters. The simplist method is to use the
Character Map window. On most recent versions of window
s you can get this by
clicking Start then Accessories then System Tools and finally Character Map.
This will open a window that displays all the characters that it is possible to
create. To enter an accented character, you can click on it's symbol then cli
ck
Select and Copy. Now you can simply paste it into the text where you want it. For
example I just "the e with two dots", I click select, then copy and now I go back to
this window and click Edit Paste ë . Do you see the "e with two dots"?

This is ok if
you only need one or two, but it's pretty slow. A faster method is also
shown on the character map window. Look at the bottom right corner of the
character map screen. Each time you click on a character you will see the
"keystrokes" that will cause this ch
aracter to display. In the case of the "e with
two dots" it says Keystroke: Alt+0235. This means that you should hold down the
key labeled Alt on the right hand side of the space bar and while you are holding
it down, press the keys 0 2 3 5 on the numeric
keypad on the far right hand side
of your keyboard. When you release the Alt key, the character that you want will
appear. You can do this without leaving the window where you are typing. I will
do this right now ë . See, it works!


Some people like to ma
ke a small table of the special characters that they
usually use and paste it up on their monitor for easy reference.


Finally, if you type a lot in languages other than English, there are special
programs that can remap some of the keys on your keyboard
directly to the
special characters that are used frequently in that language. And optionally you
can get keyboards with markings for these special characters. For example in
Ottawa Canada where most offices work in both English and French, it is quite
comm
on to find computers with "bilingual" keyboards. But as an English speaker
(and writer) I find such alternate keyboards extremely awkward. I suppose that
makes me a little linguistically chauvanist! ... grin ...


Bill.



#1.08

Languages support in WINDOW
S

2004
-
02
-
22 20:10:44
-

igr


Windows 2000 and Windows XP includes support for 126 and 136 different
locales, respectively, ensuring that users around the world can configure their
systems to use the correct formats for their own language and region.



20

P
lease see:


Language support


And in your control panel see Language Options.



#1.09

Yes, Sheila, accents are necessary

2004
-
02
-
22 20:04:40
-

igr


You have to set up your operating system to allow a Spanish keyboard and learn
the corresponding keys

in your keyboard. This action can be easily done if you
have Windows XP. What OS do you have installed in your PC?


#1.10

Thank you

2004
-
02
-
24 14:35:12
-

spage333


To my two Advisors:


I appreciate every bit of advice you give me. Please do not ask q
uestions about
this computer I am now using. It does not belong to me and I have no power to
set or change it. My first problem is that I was accustomed to the language
system of the computers at the university and could write in other languages with
accen
ts included. Secondly, (and this is chronic), I am somewhat mechanically
and electronically challenged.


The last sentence was another test. I want to see if the "challenged" expression
translates.


Sheila


#1.11

Open source multilingual portal software


2004
-
02
-
10 02:11:21
-

billpage



21

Aynisiwi is what is known as a "forum" or "discussion group" web application
program. These days such forum software is often embedded in a more complex
web
-
based system called a "portal". A portal both presents informat
ion to the
user as well as inviting the user to respond in terms of comments or by
submitting documents, pictures and other content of their own to share with other
users. Such portals are also referred to as "content management systems". Many
of these por
tals are designed to allow this interaction to take place in several
different languages. A good example of such software is Plone.


Plone is also an example of what is called "open source" or "free software". The
word "free" is used here in the sense of
"freedom" and not necessarily meaning
"at no cost". The licensing conditions of this software is such that anyone who
has access to the Internet is free to download the complete program including
the original source code, and is free to modify this program

in anyway they feel
inclined to PROVIDED that they agree to make all and any such changes
available, in turn, to all other people. A company or individual is also free to
package the program and sell this a commercial product (under the same license
condi
tions) and also to sell services in support of this product.


Typically open source programs are programmed by volunteers who freely
contribute their own time and effort in much the same way that an artist might
prepare a work of art. Except of course thi
s is a work of art that from its very
inception is intended to be shared fully with others. In some cases volunteer
effort is deliberately supplimented by commerical interests or customers who
agree to abid by the open source license conditions. It seems t
hat the
traditionally closed commercial software businesses (such as IBM) are quickly
learning that the value of the talent invested a given open source program is a
resource that out weighs the lack of control that they encur as a result of the
open sourc
e licensing conditions.


It is remarkable that in recent years a large number of very useful and high
quality computer programs have become available as open source. In particular
the entirely free Linux operating system now challenges the very product th
at has
made Microsoft the richest company in the world, Windows. And another open
source program called OpenOffice now challenges the Microsoft Office suite of
programs that is very widely used in business and government applications.


Finally, I would li
ke to mention that a lot of the sophisticated web
-
based
applications that are now becoming available are made possible by the
international standards organization known as W3C. In particular there is an
emerging new standard called I18N. Web
-
applications s
uch as Plone have
adopted this standard as a way to organize and present information in a fully
multilingual web content. Automatic translation systems have not yet become a
part of the I18N environment, but it is clear by the current experiment with

22

Aynis
iwi that it is likely that they will.


#1.12

Physics and Mathematics

2004
-
02
-
10 00:25:13
-

billpage


The invitation to participate in this forum asks me to post messages on any
subject that suits my interests. Apparently the idea is simpley to "play" an
d
experiment with Aynisiwi in order to exercise it's ability to translate. In doing this,
we may happen to meet other people
-

even people speaking other languages
-

who share some of our interests. So here is the first of several topics that I would
like
to initiate. Some of what I write will be quite technical. If you do not
understand something (or perhaps, especially if you do not understand
something) please feel free to ask me. Even if your background is not physics or
mathematics, I will be happy to
reply as best as I can.


The things that I want to talk about here include: general relativity, unified field
theories, quantum mechanics, differential geometry, topology, category theory,
tensor analysis, differential forms and computer algebra systems.
I have no idea
if I will have the time to mention all of these things in detail but one has to start
some where!


#1.13

quantum mechanics

2004
-
02
-
10 01:25:22
-

billpage


Quantum mechanics is a whole different story from general relativity. In fact one
o
f the most famous unsolved problems of twentith century physics is how to
reconcile these two theories (or even if they are reconcilable at all). Quantum
mechanics does not describe the world in terms of geometry. Rather nature, at a
very fundamental level
, is presented entirely in terms of waves. But these are
waves of a very peculiar nature. The amplitude or intensity of the waves must be
described by a complex number at each point in space. And the square of the
absolute value of that number represents t
he probability that a particle will be
found at that point in space. To really understand this, you must first understand
what is a complex number and then you have to abandon the idea of particles
and trajectories as such. We can no longer say exactly how

something moves.
Instead we find that it is the ensemble of all the possible ways that something
can move that will determine in the end how likely it is that we might find, say a
particle at location x at time t when we chose to measure it. Still, for al
l practical
purposes, the things that we can measure and how they change over time are
very accurately predicted by this new theory.


23


Although quantum mechanics is very successful as a method of calculation, it
presents many conceptual metaphysical proble
ms if we attempt to compare it to
other physical theories such as electromagnetism, or more specifically
electrodynamics, and general relativity. It seems that in quantum mechanics one
is largely forced to abandon one of the fundamental and intuitive featu
res
physical reality
-

in fact the very existence of independent objective reality itself!


#1.14

Unified Field Theory

2004
-
02
-
10 01:07:52
-

billpage


As well recognized by it's inventor, General relativity had two basic flaws. It was
incomplete in two

very important ways. One was that it did not account for the
basic properties of matter and energy. These properties had to be added
externally. Einstein liked to say that part of the theory was marble (represented
by the absolute properties of geometry)
and part of the theory was wood (only a
temporary way to represent matter and energy).


The other basic flaw of the theory was that, although it very accurately
represented gravitation, it did not include any of the other known forces of nature.
In partic
ular it did not include any of the features of the theory of
electromagnetism. It seemed quite unsatisfactory to have two very different
theories about nature since it was not clear at all how something should behave if
it was simultaneously subject to bot
h an gravitational and an electric field.


Up until the middle of the twentith century, Einstein and may other theoretical
physicists tried to develop a natural extension of general relativity that would
address either or both of these problems. Unfortuna
tely all of these attempts
were eventually rejected. The idea of a unified field theory is still very much alive
today in a new type of theory called "string theory". But new idea is not what I
would like to talk about here. Rather I would like to describe

some of the new
ideas about the older and original type of unified field theory.



#1.15

General Relativity

2004
-
02
-
10 00:50:39
-

Puesto por: billpage


Albert Einstein's theory of general relativity is fundamentally a theory about the
nature of gravi
tation. Because the effects of gravitation are largely felt in terms of

24

the way things move, general relativity also must be a theory about motion. In
order to describe motion, it is necessary to deal with both space and time. One of
Einstein's major contr
ibutions to physics was to realize that both space and time
can be treated in a uniform geometric manner
-

spacetime. Furthermore this
spacetime need not be flat like the drawings that one might make on an ordinary
piece of paper. For example we might draw

a graph with the position coordinate
on one axis and the time coordinate on another so that a curve might represent
the motion of a particle. Instead of a flat paper, we might also choose to draw our
graph on a curved space like the surface of a sphere or

on a surface like the
shape of a saddle.


Or we might attempt to describe the motion of a particle in a space which is also
twisted as well as curved. For example, think of making a drawing on a table
cloth. Now, take hold of the table cloth with the pal
m of your hand and twist it.
Part of the shape of the curve will be due to it's shape when the cloth was flat
and part of it will be due to way in which we twisted the cloth.


The idea of what constitutes the straightest or shortest line joining two point
s in
the geometry of such a curved and/or twisted surface is more complex than on a
simple flat surface. Remarkably, it turns out that the mathematics that one needs
to describe this idea is exactly the same as what is needed in physics to describe
the mot
ion of particles in a gravitational field!


#1.16

Special Relativity

2004
-
02
-
10 02:44:56
-

billpage


Before general relativity, there was special relativity. Special relativity was
developed by Einstein specifically to account for a peculiar property of

electromagnetic radiation (light, radio waves, etc.) Although such waves were
very well described by the theory of electromagnetism developed in the previous
century by Maxwell, new measurements at the turn of the century of the speed of
propagation of li
ght signals indicated that it's speed in all directions was
independent of the motion of the Earth in space. The Earth moves at an
appreciable rate around the Sun (about 30 km/sec), yet if electromagnetic waves
travelled in space in a manner similar to the

way waves travel in water (or sound
waves in air) then there should be an easily measurable difference in the
apparent speed of light travelling in the same direction as the Earth through
space compared to when it is travelling against the motion of the E
arth. It was a
big surprize to physicists at that time that it turned out to be the same!


Einstein realized that one way to explain this is if both space and time were

25

described in a single unified spacetime where the concept of distance is given by
a mo
dified Pythagorus Theorem of the form


ds^2 = dx^2 + dy^2 + dz^2
-

c^2 dt^2


This says that a small change in the x
-
direction (dx), say a motion of 1 cm.,
coupled with a small change in the y
-
direction, a small change in the z
-
direction
AND a small chang
e in the "time", say 1 second later in time, results is new kind
of small change in combined spacetime (ds). The letter c in the above formula
represent the speed of light. The number is a very big number, so the effect of
this new kind of spacetime geomet
ry only becomes important when things move
at a very high speed near the speed of light (300,000 km/sec).


#1.17

Automatic translation potential in knowledge sharing

2004
-
02
-
06 19:04:14
-

igr


Sorry to see, Gladys, that your topic presentation in Germa
n has caused panic in
Atamiri; the translation is broken; but I am sure this deficiency will be soon
overcome. Nevertheless, the topic is very important and I hope that it will deserve
replies. It is true, there are also a limit to what we can achieve with

this
technology, but we should try to explore as much as possible the given
opportunities of multilingual communication. As a matter of fact, this forum is a
first challenge that we can evaluate in real environment.



#1.18

UNESCO's policy for the use of

languages

2004
-
01
-
12 19:58:31
-

igr


The promotion of access to information in the public domain in a balanced use of
languages was approved by the General Conference 3 at its 29th session as a
basic element of this strategy and was included in its pro
gramme planning for the
1998
-
1999 biennium.


It was further emphasized in 29 C/Resolution 38 inviting the Director
-
General to
"lend fresh impetus … to linguistic diversity at all levels of education and to
multilingualism in educational curricula and to a
ssist in the further development of
educational services in Member States in indigenous and minority languages".




26

#1.19

UNESCO's Universal Declaration on Cultural Diversity

2004
-
02
-
07 13:12:53
-

phector


UNESCO's Universal Declaration on Cultural Dive
rsity was unanimously adopted
by its Member States. The Declaration aims both to preserve cultural diversity as
a living, and thus renewable treasure that must not be perceived as being
unchanging heritage but as a process guaranteeing the survival of huma
nity; and
to prevent segregation and fundamentalism which, in the name of cultural
differences, would sanctify those differences and so counter the message of the
Universal Declaration of Human Rights.


The Declaration specifically addresses the issue of
language and multilingualism
in a number of its Articles (5 and 6) and in the associated Action Plan.


The full text of the declaration is available online in Arabic, Chinese, English,
French, Spanish and Russian at this website


#1.20

RE: UNESCO's polic
y for the use of languages


Fecha: 2004
-
02
-
07 12:52:29
-

phector


The General Conference of UNESCO, recognizing the importance of promoting
multilingualism and equitable access to information and knowledge, especially in
the public domain, and reiterat
ing its conviction that UNESCO should have a
leading role in encouraging access to information for all, multilingualism and
cultural diversity on the global information networks, adopted the
Recommendation concerning the Promotion and Use of Multilingualis
m and
Universal Access to Cyberspace at its 32nd session (30 September


17 October
2003).


Through this Recommendation, which proposes measures fostering universal
access to digital resources and services, and facilitating the preservation of their
cultu
ral and language diversity, UNESCO is encouraging its Member States to
support equitable and affordable access to information and to promote the
development of a multicultural information society.


See the text of this recommendation on the web in followi
ng versions:



27

English, French and Spanish


Also available in Arabic, Chinese, Russian.



#1.21

UNESCO's main goal

2004
-
01
-
12 20:01:12
-

igr


Public domain information is a global public good; without active public support
there will be under
-
provisio
n of this good. With this in mind, UNESCO's main goal
consists in redefining universal access to information in all languages in
cyberspace by encouraging (1) the development of tools (translation
mechanisms; terminology; protocols; etc.) that will facilit
ate multilingual
communication in cyberspace (2) the promotion of fair allocation of public
resources to public information providers; and (3) the promotion of access to
multilingual public domain information and knowledge.



#1.22

UNESCO Linguapax

2004
-
01
-
12 20:09:15
-

igr


UNESCO Linguapax is a UNESCO's project which aims at promoting a culture of
peace through the promotion of multilingual education and the respect for
linguistic diversity.


#1.23

Initiative B@bel

2004
-
01
-
12 20:03:03
-

igr


The p
rogramme "Initiative B@bel" proposes to do this by implementing concrete
activities at national and international levels, with the objective to develop
multilingualism on the information networks and to encourage full partnership
between governments, indus
try and civil society.


#2.01

Hypertext translation
-

handling of HTML markers

2004
-
04
-
20 19:07:19
-

igr


28


As you probably have noticed, some of the messages contain HTML markers, for
example: See:ATAMIRI The nice feature is that the translated text hand
les those
markers in a syntactically correct way, for example: This small adjective is in the
source text before the noun, but not necessarily so in the translated text. This
examples show that our translator engine could be very easily adapted for
webpage

translation.



#2.02

Sample text from a Linux Manual

2004
-
05
-
03 17:03:27
-

igr


NAME

alloca
-

memory allocator


SYNOPSIS

#include


void *alloca( size_t size);


DESCRIPTION

The alloca function allocates size bytes of space in the stack frame of t
he caller.
This temporary space is automatically freed on return.


RETURN VALUES

The alloca function returns a pointer to the beginning of the allocated space. If
the allocation failed, a NULL pointer is returned.


#2.03

Mixed language messages can not
be translated completely

2004
-
02
-
09 14:12:19
-

igr


David: thanks for your message. Please don't forget that you have to keep the
whole text in the language that you indicated for this text, including titles. You can
try with a short text taken from suc
h a UNIX manual; the missing terms will be
included in our dictionary after few days. Then you can evaluate the translation
quality. Regards, IGR


29


#2.04

Another way to proceed with translation field
-
testing

2004
-
02
-
04 21:25:48
-

igr


You can download fr
om ATAMIRI the multilingual messenger Qopuchawi. If you
open two windows in your screen, you can simulate a bilingual dialog and verify
the translation quality in various language directions.


#2.05

test

2004
-
02
-
04 08:15:35
-

Vaneeta


I'm just testing t
his product to see the effectiveness of it.


#2.06

RE: test

2004
-
02
-
04 09:17:37
-

igr


Good, well come! Please write any kind of sentence which normally would
appear in a forum. One thing that our translator engine is not able to do, is to
guess words
when you write them wrong, for example, «iamjust» in your test
sentence. Don't hesitate to write your comments.



#2.07

Source texts in English

2004
-
01
-
13 09:41:29
-

igr


Please post your source texts in English as new contribution to this topic.


#2.0
8

Trial and error

2004
-
03
-
23 14:04:36
-

spage333


Many words in English have confusing double meanings. Look at this for

30

example:


Time flies like an arrow.

Fruit flies like a banana.


These are examples of mistakes in English punctuation and grammar
. What
happens in translation?


"Well damn, I ain't having no luck finding what I'm looking for. Might as well give
up." (overheard)

"You get a chance to see first hand what is actually preventing you in the
physical, for example, your beliefs from exper
iencing what you desire." (rising
woman
-

lifestyles, 2004)


#2.09

RE: Trial and error

2004
-
03
-
30 12:01:17
-

juanal


The text truncation problem on this and similar messages has been solved. This
enables ATAMIRI to receive the source text exactly as it
is posted.


#2.10

RE: Simply formal English

2004
-
03
-
14 15:09:35
-

spage333


from "Time for a Rhyme" by Brimax Books, Cambridge UK, 1975


Mary had a little lamb,

It's fleece was white as snow;

And everywhere that Mary went

The lamb was sure to go


It followed her to school one day,

That was against the rule;

It made the children laugh and play


31

To see a lamb at school.


And so the teacher turned him out,

But still it lingered near,

And waited patiently about

Till Mary did appear.


Why does th
e lamb love Mary so?

The eager children cry;

Why, Mary loves the lamb, you know,

The teacher did reply.


#2.11

T. S. Elliot

2004
-
03
-
02 14:04:04
-

spage333


"Where is the wisdom we have lost in knowledge . . . and the knowledge we have
lost in inform
ation?"


#2.12

Bukowsky

2004
-
02
-
09 11:43:17
-

arcangel


YOUNG IN NEW ORLEANS

Charles Bukowski


starving there, sitting around the bars,

and at night walking the streets for

hours,

the moonlight always seemed fake

to me, maybe it was,

and in the F
rench Quarter I watched

the horses and buggies going by,

everybody sitting high in the open


32

carriages, the black driver, and in

back the man and the woman,

usually young and always white.

and I was always white.

and hardly charmed by the

world.

Ne
w Orleans was a place to

hide.

I could piss away my life,

unmolested.

except for the rats.

the rats in my dark small room

very much resented sharing it

with me.

they were large and fearless

and stared at me with eyes

that spoke

an unblinking

de
ath.


women were beyond me.

they saw something

depraved.

there was one waitress

a little older than

I, she rather smiled,

lingered when she

brought my

coffee.


that was plenty for

me, that was

enough.


33


#2.13

Handling of syntactical homographs

2004
-
01
-
13 10:48:31
-

igr


She can open the can with a can opener.


#2.14

Verification of correct syntax in translations

2004
-
01
-
13 09:58:55
-

igr1


Positions of noun and adjective: An expensive project is always difficult. A very
expensive project i
s unworkable and a huge expensive project is totally
impracticable. On the other hand, a good low priced project can be very
interesting.


Position of the direct object: Our institution will finance a good low priced project.



#2.15

Invitation to partic
ipate in Atamiri's field
-
testing

2004
-
01
-
13 08:49:21
-

igr


All B@bel Forum participants are invited to post short texts in any of the five
available languages in order to test the capabilities of our translator engine.
Periodically we will observe tran
slation quality of the posted texts, if there is an
improvement during this experimental operation. The posted texts will be
organized by source languages. Suggestions are welcome!



#2.16

Which is the address?

2004
-
02
-
27 09:18:45
-

igr


Which is the
address that you said you were including?


#3.01


34

English speakers

2004
-
01
-
13 19:46:17
-

igr


Here we introduce our self those who speak English.


#3.02

Mission Titicaco
-

founding a global knowledge revolution

2004
-
05
-
14 08:12:19
-

titicacoki


Our pu
rpose is to promote the development of communication and understanding
in the world through providing existing schools in Sweden and developing
countries with access to an educational internet community
-

Titicaco
Communicaring School.


Our vision is to c
reate a knowledge revolution.


The best way to keep people poor and without power is to keep them un
-
educated. The internet changes that foundation and is also a good channel for
providing education. There is nothing that can stop the ICT development in
d
eveloping countries and the hunger for knowledge and tools to fight poverty
throughout the world. Mission Titicaco wish to accelerate that process.


What do you all think on this? Is this a good idea?


Kind regards

Ivan Luksic Daza

General Secretary of

Mission Titicaco


#3.03

Sheila Page
-

Alberta, Canada

2004
-
02
-
22 15:08:12
-

spage333


Good Day my Friends,


I am Sheila Page
-

Bill Page's sister. My most recent education involves a
Bachelor's degree in Humanities at the University of Lethbridge here

in Alberta

35

Canada.


About three years ago I became aware of this project through my brother who
urged me to use my knowledge if linguistics, anthropology as well as my writing
skills to research and document some of the interesting aspects of this unique

translation system. Now is the time for action and I look forward to the beginning
of my special project in May.


In the meantime I will post some articles in French and English. Perhaps we can
induce more discussion to happen.


Until later, Sheila


#3.
04

Bernard Lietaer

2004
-
02
-
12 14:11:22
-

lietaer


I work in five languages, so am curious what the state of the art is in the
automatic translation domain.



#3.05

Bill Page

2004
-
02
-
06 16:52:22
-

billpage


Hello.


My name is Bill Page. I have a lon
g interest in Ivan Guzman's work both on the
Atamiri language translation system and also on the particular features of the
Aymara language.


My background is in computer systems, mathematics and physics. I currently
work under contract with the Canadian
government on defense
-
related scientific
and technical information systems.


It is my personal wish to assist Ivan to more fully develop this system.



36

Regards,

Bill


#3.06

RE: Página de Factura es un bonito nombre

2004
-
02
-
22 23:21:29
-

billpage


Iva
n,


You are right, the translation does sound funny. :
-
) And I don't mind at all if you
choose to leave it as if for now. I can imagine that sometimes the translation
system cannot easy tell a personal name from some other part of speech
-

especially if i
t can be interpreted in another syntactically correct way.


In these days of the Internet, the choice of name can make a big difference to
how easy or how difficult it is for other people to find you. So far it seems that 'bill
page' is a good name for hi
ding since if you use any standard Internet search
engine, you get many many hits more of the nature of "page of bills" then
anything I have done!


Cheers,

Bill Page.


#3.07

RE: RE: Página de Factura es un bonito nombre

2004
-
02
-
24 14:14:01
-

spage333


Hi!


My suggestion as to the "Factura" problem would be that Bill call himself William
(his formal name)even though he's been "Bill" or "Billy" all his life. We can not do
anything about the "Pagina" however. It is interesting and also funny sometimes
wh
at things can appear, somewhat like speaking to someone who does not know
your language too well
-

their words do not always accurately represent their
meaning. In addition, it is not easy these days for us English
-
speakers to resist
using confusing jargon

not easily translatable into other languages. Many people
do not have the training in formal English, or the vocabulary for that matter, that
is required to make speech or text readily comprehendable.


37


What is this "religioso" confusion? I did not sense
that. Maybe I understood the
context in which it was being used but it requires another examination.


Sheila

#3.08

RE: RE: RE: Página de Factura es un bonito nombre

2004
-
02
-
29 00:03:40
-

billpage


Well, I am certainly confusing Atamiri when I use so
me language other than what
I told Atamiri I would use. Perhaps there needs to be some way to tell Atamiri not
to attempt a translation at all, like putting some special kind of quotation marks
around some text that means please don't translate.


About th
e "religious" confusion: with my almost complete lack of vocabulary in
Spanish the putatively Spanish phrase: "Facturar hermana de la Página"
sounded rather like the name of a Monk or some ancient Catholic person or
something like that.


But you are right
, in many circumstances I probably should use my real name
William rather than the more common form in English. In fact just last September
I was stopped at the airport in Paris because the name on my air ticket "Bill
Page" did not match the name on my pas
sport "William Page". It took what
seemed like a long time for the agent to check with several of her colleagues
before they eventually agreed that it was in fact the same name. I was quite
worried that I was going to miss my plane!



#3.09

RE: RE: RE: Pá
gina de Factura es un bonito nombre

2004
-
02
-
24 17:58:59
-

spage333


Spelling mistake alert!


My "comprehendable" should have been "comprehensible". Does that compute?


Sheila



38

#4.01

Posting disappears

2004
-
04
-
04 17:03:59
-

spage333


Ivan,


I have
been wondering why sometimes after a message has been posted it may
appear for a short time then disappear without a trace. Since I started using this
system I have seen it happen three or four times. Only one of the messages was
mine but still I am curiou
s as to what is happening.


Sheila


#4.02

RE: Posting disappears

2004
-
04
-
06 17:22:52
-

igr


Yes Sheila, I have deleted three postings which were not in the right place from
the point of view how to organize the topic tree structure of the forums. I hop
e it
is acceptable in this experimental environment.

Iván



#4.03

Text using HTML tags and it works!

2004
-
01
-
26 19:44:18
-

juanal


You may now post text with HTML tags.


#4.04

problem with HTML <a href ...> tags in translated text

2004
-
02
-
10 02:17:4
0
-

billpage


It seems that if I insert a link like this in an English language message and then
ask for it to be translated to, say Spanish, then the format of the link is corrupted
by Aynisiwi. Would it be difficult to fix this problem?


39


Thanks.


#4.05

Questions and comments about Aynisiwi's functionality

2004
-
01
-
16 13:58:31
-

igr


As "new contribution" here you can post your questions and comments about
Aynisiwi's functional characteristics.


#4.06

Aynisiwi is missing many useful feature common to
other web
-
based discussion
sof

2004
-
03
-
27 23:29:25
-

billpage


Aynisiwi has the wonderful feature of integrated translation of messages, but it
lacks many of the convenient features that one would expect from web
-
based
discussion software. I am afraid t
hat these limitations are having a negative effect
on the type and quantity of interactions that we have seen so far here.


Examples of important missing features include:


1) Ability to edit a message after it is submited

2) Ability to easily find new
message since the last time logged in

3) Ability to view next message in thread without returning to the index

of messages.


Other very desirable missing features include:


1) Integration with email so that messages posted to Aynisiwi are

automaticall
y sent to users who have "subscribed" to a forum.

And also ability to reply by email and to "mail
-
in" to a forum

without having to use a browser to log in.


2) Ability to search for messages based on title and text from


40

the message body (and by languag
e).


It seems likely to me that extending the current Aynisiwi program to include such
features would require a considerable investment in programming time that
would probably best be applied to the translation capabilities themselves rather
than in desig
ning a better user interface.


There are several freely available forum software packages which include all of
these features, for example PhpBB and PhpNuke" and of courese Plone that I
have mentioned before. I think it would be very good to have the Ayni
siwi
translation features incorporated into one of these more full featured web
-
based
discussion tools.


Regards,

Bill Page.


#4.07

Excuse me, I have edited the title of your message

2004
-
02
-
09 12:11:21
-

igr


Excuse me, Patxi, but I have edited the t
itle of your message, because you left it
in English, when your text was written in Spanish. The title of a reply can be
changed as you wish. I hope that with additional lexicographical work the
translation of this your comment will look much better.

Yes,

the question of a cost
-
benefit analysis is crucial, but compared with what
other alternative?


#4.08

Editing and deletion of posted texts

2004
-
01
-
14 10:31:54
-

igr


Once a text has been posted it cannot be edited, therefore don't forget to revise
your
message before submiting it. The Aynisiwi administrator can edit and delete
texts, but this action is hardly ever done.


Post
-
editing of translated texts is possible, but this action will be avoided,
because we want to observe how translation quality impr
oves during operation of
this our experiment.


41


#4.09

RE: Editing and deletion of posted texts

2004
-
02
-
10 01:32:38
-

billpage


I hope that the ability to edit messages can be soon added to Aynisiwi. I am
afraid that I am almost never able to enter a mess
age with out make some
serious typing errors. In some cases, after I re
-
read something I wrote I wonder if
anyone could guess what I really meant to say! I am quite sure that this could
cause Aynisiwi some serious problems with translation. Without the abl
ity to
correct my mistakes, it makes be rather afraid to click the "submit" button...



#4.10

Aynisiwi's tree structure
-

The use of (+)

2004
-
01
-
13 12:44:56
-

igr


In Aynisiwi, the data base of posted messages is organized in a tree structure;
each for
um contains topics; topics contain contributions; contributions may
contain comments, and so on. By clicking the (+) icon at the left side of a title you
can unfold the branch and display its lower level branches. Reversely with the (
-
)
icon, you can fold
together the branch.


Before you post a message, look at the right upper corner of your window to be
sure at what level you want to locate your text. If necessary, you may go back to
the prior window using the corresponding button of your browser. Please,

avoid
posting new topics when it is not the case, instead you should post a contribution
to the topic that is currently displayed. In this way we keep an optimal overview of
our data base contents.


#4.11

Who has posted a text, author and language

2004
-
01
-
13 18:41:49
-

igr


The person who posts a message is not always its author. It is recommended
that the authorship of a posted text is acknowledged either in the title or at the
bottom of the text.


When you post a message it is very important than you

indicate the language of

42

the text, otherwise its translations will be full of stars (*) which means that it is an
unknown word in the language you erroneously gave.


#4.12

The display of titles

2004
-
01
-
13 16:33:08
-

igr


When you come in the webforo, t
he webpage displays the titles of the first 8
topics of the forum. If there are more than 8, you can click at the
-
> icon at the
right upper corner in order to display the next 8 titles. The icon appears only if the
number of titles is greater than 8. The
titles are ordered by date and time, the
latest postings are listed first.



#4.13

Topic titles and subtitles

2004
-
01
-
14 10:09:29
-

igr


When you post a new topic, the title field is blank so that you can write your own
topic title. When you write a l
ower level text, for example, a contribution to a
topic, the subtitle field contains per default the standard RE, but we recommend
that you erase it and write your own subtitle.


43

Annex 3

List of missing terms (Source text: English)


31/05/04 12:07

Lista de Términos Faltantes Hoja: 1


Documento: C:
\
...
\
E
\
PostingsE.E.txt


Idm Núm Término Inglés


D 1 As well as w
ell


D 1 But still but still


D 1 Catholic catholic


D 2 Chinese Chinese


D 1 Declaration of Human Rights declaration of human righ


D 1 Edit edit


D 1 Field
-
test posting field
-
test posting


D 1 Hi

hi


D 1 Hypertext hypertext


D 1 ICT ICT


D 3 In particular

in particular


D 1 Linguists linguist


D 1 Map map


D 1 Post
-
editing post
-
editing



D 2 Posting posting


D 1 Postings posting


D 1 Public domain public domain


D 1
Reversely reversely


D 1 Rhyme rhyme


D 1 School school


D 1 So far

so far


D 1 Source texts source text


D 1 Thanks thanks


D 2 accented characters ac
cented character


D 1 accented character accented character


D 1 account account


D 1 allocator allocator



D 1 amplitude amplitude


D 2 anything but anything but


D 1 appreciable appreciable


D 1 as such

as such


D 1 as well as well


D 1 as yet as yet


D 1 asks

asks


D 1 at the bottom of at the bottom of


44


D 1 authorship authorship


D 1 awkward awkward



D 1 businesses businesses


D 1 but still but still


D 1 caller caller


D 1

carriages carriage


D 1 charmed charm


D 1 check out check out


D 1 chronic

chronic


D 2 comprehendable comprehendable


D 1 cost
-
benefit analysis cost
-
benefit analysis


D 1 curricula
curricula


D 1 depraved depraved


D 1 deserve deserve


45


IGRAL


31/05/04 12:07 Lista de Términos Faltantes Hoja: 2


Documento: C:
\
...
\
E
\
PostingsE.E.txt


Idm Núm Término Inglés


D 2 directions

directions


D 3 edit edit


D 1 effect effect


D 1 effects e
ffect


D 3 email email


D 1 embedded embedded


D 1 envisioned envision



D 1 erroneously erroneously


D 1 every bit every bit


D 1 excerpt excerpt


D 1 externall
y externally


D 1 fair fair


D 1 fake fake


D 1 fearless

fearless


D 2 field
-
testing field
-
testing


D 1 flaw flaw


D 1 flaws flaw



D 1 fleece fleece


D 1 followed followed


D 1 free software free software


D
2 freely freely


D 1 fundamentalism fundamentalism


D 3 gravitation gravitation


D 2 gravitational

gravitational


D 1 hesitate hesitate


D 1 homographs homograph


D 1 horses

horses


D 1 hypothesis hypothesis


D 1 impracticable impracticable


D 1 in detail in detail



D 2 in general in general


D 1 in turn in turn


D 1 information society information society


D 1 integra
ted integrated


D 1 key key


D 2 km km


D 1 language divide

language divide


46


D 1 lead one lead on


D 1 lifestyles lifestyle


D 2 lingered linger



D 1 linguistically linguistically


D 2 link link


D 1 links link


D

2 lot lot


D 2 map map


D 1 markings marking


D 1 measurable

measurable


D 1 mechanisms mechanism


D 1 mentioned before mentioned before


47


IGRAL


31/05/04 12:07 Lista de Términos Faltantes Hoja: 3


Documento: C:
\