New to Typophile? Accounts are free, and easy to set up.
Create an account
Typophile RSS | More Feeds
I like it. Now for the glyphs.
Well, first off I'd like to congratulate Mr. Kumar on the selection of his design and its excellent appearance. I'd like to know if anyone knows if the Unicode consortium has a temporary unicode we can start using now, or whether the unicode is the last step and we all have to wait...
I'm checking with the MS Unicode folks, will let you know.
As to encoding, M. Everson announced that he is preparing a proposal already.
Now to the glyph. What I have seen so far very much reminds me of the Euro-glyph mess we had some years before. Well, the fontist community sorted it out rather quickly and trashed the initial “official” version.
What is that spell laying on introduction of new monetary signs? A government honours a “design” which again is of much frankensteinian technocratic inventiveness but next to nothing of typographic feeling. A bastard base glyph out of Devanagari RA and of Latin R! A clash of typographies – or a great moment of globalisation?
It’s quite obvious how the glyph will look like in Arial or Comic Sans. But how … in Times, Bodoni, Garamond? The nature of the two horizontal strokes still leaves much to be determined. I’m curious to what life this creature will grow.
Andreas: It’s quite obvious how the glyph will look like in Arial or Comic Sans. But how … in Times, Bodoni, Garamond? The nature of the two horizontal strokes still leaves much to be determined.
I suspect that the top bar, which in a Devanagari type should probably correspond to the head stroke, in European type styles will conform to the weight of the bar of the euro, yen and sterling.
Frankly, it's not a very happy shape, and yet more evidence of what I stated when the euro symbol arrived on the scene: currency symbols should not be invented by designers, they should be invented by market grocers writing with jiffy markers on cardboard boxes. As a more general rule, people inventing new signs for writing systems should write them, not treat them as architectural projects. And they should certainly be aware that a shape that works nicely in a monoline, sans serif form isn't necessarily going to work well when subjected to higher contrast and variable stroke axis.
Anyhoo... here's a quick run at a Brill rupee symbol.
David, Unicode doesn't do temporary codes, so this would need to go through the whole Unicode and ISO 10646 proposal and balloting processes before it gets a standard code assigned. Since the symbol has official backing from the Indian government, I wouldn't anticipate any problems in that process, but it will still take time.
Michael Kaplan has suggested, however, that the encoding could be handled as a reference glyph change for U+20A8. That seems to me a sensible idea, since it would not require any updates to system and application software, and only updates to some existing fonts.
Does Mr. Kumar have an alternate design for drawing heavy fonts that won’t result in the upper counter turning into a black blob?
The fact that the original drawing could be either a modified Latin letter R, or a modified glyph from the Devanagari script, suggests that designers not conscious of both contributions to its shape might face pitfalls adding it to typefaces.
Good question, James. In the other two-bar currency symbols, one has the option to reduce to one bar in heavy designs and know that the symbol will still be recognisable. [In TTF one can also use hints to collapse two bars to a single bar at small sizes in low resolutions.] But in the new rupee symbol one of the bars also functions as a head stroke, so there's no way to remove the lower bar without risking confusion with the Devanagari letter र. I think the solution in a heavy design would probably be to drop the lower bar so that it crosses the stroke reversal going into the leg.
I'd say this is pushing legibility of the symbol.
The committees that decide on new currency symbols never seem to have someone like me on board, whose first instinct in any situation is to ask ‘What are the ways in which this could all go horribly wrong?’
With the cross bars on currency symbols, my impression is that generally they are thinner than any normal horizontal—that being part of their differentiation as currency symbols. Thus, not bolding them so much in an otherwise bold weight would be an option.
Your observation about other two-bar currency symbols is correct, Bill. But in this case the top bar isn't crossing a symbol as in the euro or yen, it is forming the head stroke. It's more part of the structure than something crosses the structure.
I suppose one could apply the analysis that this is not, in fact, a two-bar symbol, but a single cross bar applied to a shape that includes a head stroke like र. In that case, a rational argument can be made for allowing the bars to have different weights, with the cross bar being lighter than the head stroke. But rational arguments do not guarantee pleasant glyphs.
Is Brill an upcoming project of yours? Looks lovely—and know I’m get off track here, but I would like to know when we can see more gems from you/Ross/Tiro? Plantagenet Novus?—what happened to than one?
I happened to be working on monetary symbols tonite . . .
Screw it . . . it's uni20A8 until further notice. It's a nice symbol but it looks like it would be hard to draw (write). Better than the crummy euro.
Typodermic: compare all the crossing bars of your regular and bold $, ¢, £, € and Rupee glyphs. What you show is highly inconsistent.
I think it was not a good idea of them to create a Roman-Indian bastard base glyph. This is hurting me. And as the first samples show, it does not work graphically. Because it cannot work. The glyphs will end up being of either Roman or Devanagari stock. And there will be endless complaints to fontists “your Rupee sign looks weird”.
>‘What are the ways in which this could all go horribly wrong?’
Kerning, rendering, scaling, glyph substitution, font portability and font formats have all proceeded this way. To remain consistent, "I don't know what you're talking about" or "the horse is out of the barn."
>Does Mr. Kumar have an alternate design for drawing heavy fonts that won’t result in the upper counter turning into a black blob?
As a pratical matter, who does this come into use? That is, how do office workers and designers get this added to fonts they have? What happened when the Euro symbol was adopted?
A proposal to add the character to the Unicode Standard and ISO/IEC 10646 was published yesterday. See
Here’s my ﬁrst humble take:
Good point. Fixed.
>I lean towards understanding this as a *Devanagari character* and nothing else.
Considering that it is the monetary symbol of India, neither your statement (a *Devanagari character* and nothing else?), nor your drawing, make sense. But maybe that is what 'first takes' are for?
I have two concerns about your document.
1) An ISO proposal document should not appear as a memorial to a single person who has invented a miserable glyph for a character he not invented.
2) quoting from http://std.dkuug.dk/jtc1/sc2/wg2/docs/n3862.pdf,
The shape of the currency sign has been specified as “an amalgam” of the DEVANAGARI LETTER RA, and the LATIN CAPITAL LETTER RA and it is likely that many fonts will take the Latin capital letter as the starting point for design, to harmonize with European digits and other currency signs.
Drop the whole bit. It only pretends to make assumptions manifest which are utmost questionable. This is not any good.
(Personally, I doubt that a Roman base glyph per se fits better to digit glyphs which originate from India; see my previous post ;-) cheers
– you’re right, I should have written “Devanagari (based) GLYPH”. Ought to make sense that way.
… nor your drawing, make sense
What does your drawing look like?
Drop the whole bit.
If the description is part of, or at least conforms to, Mr Kumar's design, then of course it should be included in the "specification". It gives designers a valuable clue as to what this sign is and what it ought look like. It helps avoid deviation. After all, this is what has been submitted and selected.
Andreas Stötzner @ 16.Jul.2010 4.37am: Therefore the terminations of the crossbar look the other way.
Which isn't the way it looks in Kumar's design.
Am I correct in gathering from this page that devanagari numerals do not use a head stroke?
Of course not.
As has been said before many times, the way and mode these “designs” (Euro, Rupee) are presented is misleading. It is plainly wrong to represent it as a technical drawing, cf. the Euro introduction.
The visual essence of such a sign (not ‘symbol’ btw) is its graph, the skeleton. The *rendering* of the sign’s skeleton is something else. Something very important, naturally. That is what has been ignored again.
As John Hudson repetedly remarked, a monetary sign should get “invented” rather by people who actually write it. Instead we witnes again a competion which is ruled by bloody laymenship sailing under the label “design”.
Why is WRITING not recognized being a design discipline?
I have read statements which favour the new sign being based on a Latin shape rather than on a Devanagari one. Considering the countries multiscriptive reality that has a point in terms of attempting neutrality. But if so: is the attachment of a Devanagari-headbar onto a Latin R-glyph such a good idea? The winning “designer” seems not to have bothered himself with such thoughts. Instead he was inspired by the Indian flag (!) and created an Roman-Indian bastard. If that is design, then throwing mud and mustard into one bucket is cooking. ;-)
However, we’ll sort it out. No choice.
Here's an amateur go at a didone version.
>The fact that the original drawing could be either a modified Latin letter R, or a modified glyph from the Devanagari script, suggests that designers not conscious of both contributions to its shape might face pitfalls adding it to typefaces.
I disagree. In a European-language font you make it R like, in a Devanagari font you make resemble the base characters of that particular design. Of more concern, as Andreas notes, is what this should look like in fonts supporting the other dozen or so Indian scripts in common use.
compare all the crossing bars of your regular and bold $, ¢, £, € and Rupee glyphs. What you show is highly inconsistent.
Well, good. I'm glad you noticed. Yeah, of course the crossbars are inconsistent. I never even considered having evenly weighted crossbars as being something that would be desirable. Why would I want that? Why would that be something anyone would want to attain unless I were entering some kind of even crossbar font contest. The crossbars don't need to relate to one another in currency symbols. They can, and it certainly looks fine in light weights. But what's the advantage? They aren't usually seen together. Numerals are their natural companions. A yen and Euro can have thinner crossbars, a cent and sterling can have thicker. Does it look good with numerals. Then the crossbars are the correct thickness. In my example, my dollar sign cross strokes happen to be short. Of course I don't want to match them to the Euro. They need to be fattened up, they need presence. The thickness is determined by the density and relation to numerals and my particular mood. Looking good with numerals trumps all. That's the most important thing. Although Ke$ha complicates things. Why, Ke$ha?
It's hard to guess how grocers will write the rupee. From my marker scrawlings, it seems more practical to increase the size of the top bowl and diminish the tail. I tried drawing about 100 of these but I am not an Indian grocer.
All that I've discovered it that the new rupee symbol is really god damn hard to draw and that I'm very glad I'm not an Indian grocer. Maybe it'll relate to the R because of the Rs. Anyway. Grocers, please get to work and figure this out for us.
… and the more concern with fonts which support Latin AND Indian scripts.
Eliason: not bad! I too feel that the two hor. strokes might have different thickness in contrastive faces.
There are, practically, going to be three different models for this symbol, based on usage.
The official version, as designed by Mr Kumar, will presumably appear on bank notes and coins.
A vernacular written version will presumably develop in India, and I suspect it will be a र, written with one of the two typical ducti that writers of Devanagari employ, with a crossbar through it. Understanding these ducti will be useful for makers of handwriting fonts (Ray, I think your construction is atypical: either the head stroke is written first as a separate stroke, or the body is written first and then the head stroke added above; in a lot of informal Devanagari handwriting, the head stroke ends up being above the body of the letter, rather than attached to it). This form will be used mainly in conjunction with Indic style numerals. It is worth noting that different Indian scripts have their own numeral symbols, and that there may be localised vernacular forms of this currency symbol that derive from the way the individual scripts are written, or something more Latin R-like may be preferred on the basis of national script neutrality. [There are at least some people in India who favour a distinct rupee symbol for each script, which could still happen alongside the official federal version.]
A Latinised form will harmonise with European fonts and will be used primarily in the context of European numerals. In language tagged text, an OTL 'locl' variant could be used to switch between a European styled and Indian styled rupee symbol.
Forrest: As a practical matter, how does this come into use? That is, how do office workers and designers get this added to fonts they have? What happened when the Euro symbol was adopted?
The euro symbol was a different beast in terms of software support, font updates, etc.. It was something that was added to existing 8-bit Latin codepages, and hence to all fonts that claimed support for those codepages, and of course to any font that one realistically wanted to license in a European market. It was a huge, huge undertaking.
As far as I know, only Apple among the major OS makers have 8-bit codepages for Indic scripts, with few fonts targeting those codepages. Microsoft's support for Indic scripts has always been Unicode-centric, and I didn't notice any IBM or other codepages for Indic scripts. So there will be a lot less pressure to update the majority of fonts, which do not support Indic scripts. The rupee symbol will probably be like the majority of currency symbols in Unicode that are only found in fonts that support particular languages/scripts, or in fonts that aim to support complete Unicode blocks.
Typodermic: The crossbars don't need to relate to one another … They aren't usually seen together.
The are. Just one sample, from http://www.ft.com/home/uk :
Yes, probably. But I argue the Indians want their new sign of economic pride getting viable on a more general scope, in the end *one* glyph in a given font should perform satisfactorily with either Latin or Indic script environment. The sophistication of handling several localized glyphs via OT features won’t work for the Essex man.
This is quite a graphic challenge. The Indian officials decided to leave the solution up to us. Maybe not the worst case.
This certainly isn’t in the same league as the euro, this isn’t a new currency, and thankfully we’ve moved beyond the 8bit code-pages. I don’t think the existing Rupee symbol is even on standard Indian keyboard layouts (could be wrong). Checking some shopping sites they seem to use “Rs” rather than the current Unicode codepoint. I don’t see that changing until fonts including the new symbol are ubiquitous. Beyond shopping sites the business and banking world will continue to use three letter currency codes, esp for second tier currencies.
Also, I wonder if the designer of this symbol is related to Dr. Anshul Kumar, who teaches at IIT, and who has a lecture series available online:
Monetary symbols born of any era have problems in typographic transition. The U.S. Dollar is no prize either in a heavy weight, particularly the double bar variety. The new Rupee is no worse than the rest. The slash lines are what mark all of them as monetary. The problem the Euro had was that it was too exact when first released. Now what has happened to it will happen to the Rupee as well. Five hundred type designers will putz with it to make it fit their own typefaces and a general feeling of what reads as the symbol will evolve and make itself known with time. At least this new Rupee does not have a construction diagram to go with it. I take it as a fun exercise and am happy it is not more styled or designed as to be impossible to adapt.
There will be rumbles of discontent until the real code point is fixed. Then we will just all go to work and draw it like we always have.
A glyph change to U+20A8 is out of the question, it turns out. That codepoint has a compatibility decomposition to R+s and, for reasons of Unicode stability agreements with other standards, that decomposition cannot be changed. This disqualifies U+20A8 from use for the new rupee symbol.
My document gives the designer because this is the result of a contest. That's part of the history of the character. I did the same thing in http://std.dkuug.dk/jtc1/sc2/wg2/docs/n3392.pdf the proposal to encode the Kazakh TENGE SIGN.
As to your second point, well, you're just wrong. Firstly, using Devanagari RA as obviously as you did in your font would be inappropriate for a Latin-based font, and moreover a lot of Indians use other scripts and feel that Hindification via Devanagari is not a good thing.
Secondly, the thing WAS IN FACT devised as an amalgam of Devanagari RA and the right half of Latin R. And so far the most successful designs for the character I have seen have taken R as the starting point. Just as the most successful EURO SIGNs take C as their starting point.
>The winning “designer” seems not to have bothered himself with such thoughts.
Please. This skeleton is a marvel. Look at all the possible approaches!
>What does your drawing look like?
I have not been hired to draw one yet. But the bars will match the other monetary signs, and so will the overall complexity of the shape.
>A glyph change to U+20A8 is out of the question...
Let me know.
I'm not sure we'll have to add it to all of our fonts, but we'll do Ravie first and check with you all to see what you think.
Here's a possible monkey wrench thrown into the works, or spanner in the works as the British say:
(From Theo Rosendorf on Facebook.)
@ Craig : Am I correct in gathering . . . that devanagari numerals do not use a head stroke?
That is correct.
Here's a possible monkey wrench thrown into the works ...
What does this mean:
Second violation of guideline is the symbol itself, as it's not applicable to standard keyboard or Unicode enable.
The crossbars don't need to relate to one another … They aren't usually seen together.
The are. Just one sample, from http://www.ft.com/home/uk :
It doesn't matter. They're not together. Even if they were on the same line, the strokes wouldn't need to match. In heavy weights, the top of a T and bottom on an L don't need to match. The crossbar of an A and a crossbar of H don't need to match. The strokes of a seven and four don't need to match. When you're adjusting weight and trying to harmonize with other glyphs, there are many options, one of which is changing the weight of strokes. A euro is dense. A sterling is less dense. By restricting crossbar thickness variation, you have one less tool in your toolbox for dealing with the problems that arise in heavier weights. On course, if you're making book and light, you have the option of unifying strokes with little consequence.
The mistake I made with my rupee was not using a similar strategy to what I used for the sterling and using a thinner middle crossbar. I think it might be sensible to look towards numerals for the rupees construction and ignoring the R altogether. The sample shown in the Times article shows a straight horizontal stroke, which you could relate to the treatment of the stroke on of your seven. Maybe not a copy of the 7 stroke but somehow related, rather than using the R stroke as a cue. Maybe the top bar could relate to the 7, 5 or, if available, a flat topped 3. The middle could relate to the middle of a 3. You can start with an R if you like but eventually, your rupee will have to get along with numerals. When I make a euro, I use the zero as a starting point because it's not going to be hanging out with letters very often.
Bill, thanks for that great link. I loved the television clip.
@ Christopher Slye - Down below are the guideline issued by Finance Ministry in regards to Indian Rupee symbol design competition.
According to guidelines : http://www.saveindianrupeesymbol.org/2010/06/indian-rupee-symbol-guideli...
. A participant can send a maximum of two entries.
. The symbol has to be in the Indian National Language Script or a visual representation.
. The symbol should be applicable to standard keyboard.
But RTI (Right To Information)had already exposed that he is Non-eligible candidate for Indian Rupee Symbol Design Competition as he had violated the Indian Rupee Symbol Design competition guidelines .
According to guidelines one candidate could send maximum two design entries but he had submitted a total of four designs.
Second violation of guideline is the symbol itself, as it's not applicable to standard keyboard or Unicode enable.(Controversial guideline)
Violations are not stopping here according to information and broadcasting minister Ambika Soni "The new rupee symbol partially resembles the Devanagri ‘Ra’ and Roman capital ‘R’ without the stem coupled with two parallel lines – in line with Finance Minister Pranab Mukherjee’s vision for a symbol which reflects and captures Indian ethos and culture."
But using roman script/English alphabet in designing the New Indian rupee symbol is also a violation of guideline according to the guideline "The symbol had to be in the Indian National Language Script".
But according to National language Script English is not included in that list till this date.
More over RTI Activist exposed : Mismanagement and violation of guidelines in “Indian Rupee Symbol” design competition.
According to RTI Documents : http://www.saveindianrupeesymbol.org/2010/06/down-below-are-rtis-filled-...
1. Non-eligible candidate was shortlisted in top five finalists. As candidate had submitted four designs.
2. One finalist was in contact with Finance Ministry and RBI prior to competition.
3. Design concept or brief was not put in front of jury along with Indian Rupee symbol design.
4. No marks or grades were allotted to selected (2644) candidates design entries, in the process of shortlisting top five finalist design entries.
5. Jury had spends less than 20 seconds on each design entries to analyze it .
6. Three jury member were absent in two days long meeting dated 29th Sept. and 30th Sept. 2009.
7. One jury member from Ministry of Culture was absent on the day of final presentation.
8. All seven jury member had never meet in this whole selection process in any given time.
9. No records are available with Finance Ministry which could indicated, how many total design entries Finance Ministry had received ?
Typodermic: It doesn't matter. They're not together. Even if they were on the same line, the strokes wouldn't need to match. In heavy weights, the top of a T and bottom on an L don't need to match. The crossbar of an A and a crossbar of H don't need to match. The strokes of a seven and four don't need to match. When you're adjusting weight and trying to harmonize with other glyphs, there are many options, one of which is changing the weight of strokes. A euro is dense. A sterling is less dense. By restricting crossbar thickness variation, you have one less tool in your toolbox for dealing with the problems that arise in heavier weights. On course, if you're making book and light, you have the option of unifying strokes with little consequence.
Thanks for enlightening me.
No problem. I'm going to wait until there's a proper Unicode location and a final design before I make more of these but it's fun to explore.
As Unicode definition for new Rupee sign will take some time, what to do with fonts under development? To include the new sign as a non-coded glyph, to replace traditional Rupee sign or to ignore it until there is an Unicode defined?
I have the same questions about Kazakh Tenge sign, except that Tenge does not replaces any glyph.
As a first approach, I thought to include new Rupee as an alternate to traditional Rupee sign. And to simply add Tenge as a non-coded glyph.