W3C home > Mailing lists > Public > public-silver@w3.org > April 2020

Clear Language Testing

From: John Foliot <john.foliot@deque.com>
Date: Thu, 2 Apr 2020 11:48:37 -0500
Message-ID: <CAKdCpxwLEG_XzARo7JggvhmADG6o_REZHmGjitPyQaERkUSB+Q@mail.gmail.com>
To: Jeanne Spellman <jspellman@spellmanconsulting.com>, Silver TF <public-silver@w3.org>
Hello All,

Prompted by Rachael's scoring proposal, I've taken a quick look at the
tests related to Clear Language, and at this time I have significant
concerns related to internationalization.

Specifically, we appear at this time to be using the following as a basic
test assumption (from:

...and echo'd in Rachael's document (
Basic Tests


   Is spelling correct? (automated)

   Is grammar correct? (automated)

   *Is active voice used? (assisted)*

   *Is a simple tense used? (assisted)*

Some quick and basic research suggests to me that this is very
"western-centric" and will not scale to all languages. Please see:

Additionally, while there are 12 major verb tenses for English related to
"time" (past, present, future - divided into four aspects: the simple,
progressive, perfect and perfect progressive), that may not always be
applicable. Chinese, for example, has no grammatical verb tenses. Other
languages, like Indonesian, express time only through adverbs — there are
no changes to the verb form. (source:

I do not claim to be a linguist or internationalization expert, but if 15
minutes of research turns up these concerns, then I must respectfully
suggest we need to stop and rethink how to make this work for *all*
languages, and not just western/romance languages.

I do not at this time have a proposal for a path forward, but at the same
time I am speaking up to say I think we are currently on the wrong path.
Testing for tense and 'voice' (active/passive) are language specific, and
not applicable everywhere.


Received on Thursday, 2 April 2020 16:49:29 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 24 March 2022 20:31:47 UTC