Re: Persian decimal separator

From: Michael \(michka\) Kaplan (
Date: Thu Nov 16 2000 - 07:52:54 EST

The only question I have here is that no one was *ever* suggesting a slash
be used for the decimal separator. The suggestion was made for the CURRENCY
separator. Since the question was asked (and all of the answers were made)
mentioning the decimal separator only, I would have to wonder whether they
were answering the actual issue that was under consideration here?

I have two acquaintances (not connected to MS) who speak Farsi, neither is a
linguistic authority or pretends to be. Both recognize the slash as a
currency separator, but one of them believes it to be inappropriate to use
on non-Farsi currency amounts and both of them admit the matter is mostly
theoretical since the Farsi currency does not currently see a need for a
currency separator (something we discussed last time, I believed). This is
hardly conclusive proof, but if non-technical native speakers can cast doubt
on an issue, then obviously there is reason to believe there may be an issue
worth investigating?

I am not saying that you or they are wrong and would never pretend to have a
true opinion here other than to say I DON'T KNOW (until and unless I live in
Iran and learn Farsi to the point where I can respond to the question!). But
I am unsure how a company (such as Microsoft, which currently lists the
slash as the currency separator for Farsi) can proceed to consider evidence
that may well be tainted by having the wrong question asked? Clearly there
was some basis for believing that the currency separator was a slash, and
there should be a compelling basis for believing that this was a mistake?


Michael Kaplan
Trigeminal Software, Inc.

----- Original Message -----
From: "Roozbeh Pournader" <>
To: "Unicode List" <>
Sent: Wednesday, November 15, 2000 3:22 PM
Subject: Persian decimal separator

> Dear All,
> Some time ago, there was a discussion here about the Persian decimal
> separator. I am posting a short report about our queries into different
> Iranian bodies. Sorry for the long and somehow formal thing, but it seems
> important to us.
> I'm still waiting for responses from Iranian Academy for Sciences (IAS),
> and Iranian Mathematical Society (IMS). I have answers from these sources:
> * Iranian Academy for Persian Language and Literature (IAPLL);
> * Iranian Standards and Industrial Research Institute (ISIRI) which is
> the national standard body;
> * Iran University Press (IUP), and Fatemi Publishing Institute (FPI),
> which are the largest and highest quality academic publishing houses
> in Iran.
> I think that IMS will answer the same as FPI, since they seem to use the
> same conventions in their books that is not published by any of these two
> houses. They certainly use the house rules when they publish with one of
> these two, but not with other houses.
> I also add our conclusions, as current representatives of HCI (Iranian
> Council of Informatics) in text encoding issues which is the responsible
> body for national computing related standards, which is transfered to it
> from ISIRI.
> 1. All sources agree that slash and decimal separator should be considered
> different.
> 2. ISIRI has a character set in their standards (ISIRI 3342, the
> rarely-used national standard) which distinguishes the two characters,
> while not distinguishing hyphen from minus or colon from division sign
> (of which the latter case is really weird). They did not give any
> special comments regarding the standard, since the standards commitee
> for the character set issues is dissolved for a long time, and the
> responsiblity was handed to the HCI. The standard shows the glyph for
> the decimal separator as described in 4. They have also another
> standard (ISIRI 2901-revised:1994) for keyboards, that distinguishes
> the two characters.
> 3. IUP and FPI already use the same publishing software that distinguishes
> these, for more than five years, and IUP has distinguished them even
> before that time. They both agree that the the sequence ONE SLASH TWO
> means 0.5 and not 1.2. They specially say this because of the need for
> clear interpretation of in-text formulas. (IAPLL sees this
> interpretion lying beyond its competence, and refered us to the IAS.)
> IUP has also published a scientific style guide which explicitly
> mentions the difference, and asks for a glyph shape described in the
> last part of the next item (I can provide you with copies of the page
> mentioning this. We also use software that distinguishes these.
> 4. All except IUP agree that the glyph shape for the decimal separator
> should be a shortened, lowered and possibly more slanted slash. But
> IUP has changed the default behaviour of the mentioned software to
> use a glyph exactly similiar to the isolated form of REH (U+0631)
> for the decimal separator. This has been the case even in their old
> books, before their adoption of computer software for publishing.
> But the IUP recommendation in this case is considered old tradition
> by others, including us, and not acceptable. (I can provide digital
> images of text produced by FPI, IUP, and ourself.)
> 5. All except IUP agree that in the case of lacking decimal separator
> in the software, a slash is the best substitute. IUP prefers the
> REH shape in all cases. FPI insisted that using the slash for both
> division and decimal separation is unbearable, and told that in the
> case of a lacking decimal separator glyph, all the text should be
> scanned for use of slash as division sumbol, and those cases
> to two dimentional fractions.
> 6. In the case of missing Persian shape for the decimal separator,
> IUP and FPI (and also us) prefer the Arabic shape over the slash.
> IUP may also prefer the Arabic shape over their REH shape, but that's
> not verified yet. IAPLL prefers the slash over the Arabic glyph.
> 7. All sources agree that for date separation, one should only use the
> slash.
> As final conclusion:
> In case of information interchange, when the character set permits,
> a decimal separator (U+066B) is certainly prefered to a slash (U+002F)
> and must be used. Computer programs should render the Persian U+066B as
> a shortened, lowered, and possibly more slanted slash; this should be
> distinguishable from the slash at the first sight (I can provide
> examples). If the Persian shape is lacking, if the text context is
> mathematical, the Arabic shape must be used. In other cases, the slash
> shape is acceptable (but will be considered illiterate or
> nonprofessional, somehow similiar to using spaces instead of zero width
> non-joiners).
> (We have not yet received enough responses to our queries about the
> thousands separator, but it seems that there will be a lot of disagreement
> about this. I can only tell that the national character set and keyboard
> standards, ISIRI 3342 and 2901 mentioned in section 2 above, distinguish
> the thousands separator from comma, apostrophe, etc. The glyph shape
> is similiar to the shape seen in the Unicode charts for U+066C).
> --roozbeh

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:15 EDT