En and Em dashes

Question

I have recently noticed that WordPad in Windows XP replaces both en and em dashes with hyphens. No other high-end ASCII symbols are touched, as far as my experiments have been able to confirm. Is this MS's latest attempt at "simplifying" orthography?

It is not so much of a problem for me these days; I'm learning TeX, which automatically inserts en and em dashes when it sees multiple hyphens. -- Simon R. Hughes

Usenet · Accepted Answer

[nq:1]It is not so much of a problem for me these days; I'm learning TeX, which automatically inserts en and em dashes when it sees multiple hyphens.[/nq]
Good man. (ITYM LaTeX.)

Usenet · Answer

Thus spake R F:
[nq:2]It is not so much of a problem for me ... inserts en and em dashes when it sees multiple hyphens.[/nq]
[nq:1]Good man. (ITYM LaTeX.)[/nq]
LaTex is the interpreter so that would be the thing that inserts the dashes. I am learning the code, which is TeX. -- Simon R. Hughes

Usenet · Answer

[nq:2]Why bother? You can write PostScript, and feed it to the convertor. Being real men isn't about doing unnecessary work,[/nq]
[nq:1]It's possible that you've missed some of the joke. An essay appeared quite a long time ago, in Playboy I ... that stuff in your head. So to a large extent it in fact was a paean to doing unnecessary work.[/nq]
Here is one version. I'm not sure if it's identical to the first one I ever saw, but I can recognise some of the phrases (especially the ones that are, or were, a bit too close to home...):

http://www.pbm.com/~lindahl/real.programmers.html

-- (Email Removed)

Usenet · Answer

Simon R. Hughes (Email Removed) writes:
[nq:1]Thus spake R F:[/nq]
[nq:2] Good man. (ITYM LaTeX.)[/nq]
[nq:1]LaTex is the interpreter so that would be the thing that inserts the dashes. I am learning the code, which is TeX.[/nq]
LaTeX is (or at least was initially) nothing more nor less than a big set of TeX macros. The "latex" compiler was simply the "tex" compiler with "latex.tex" preloaded. LaTeX was Leslie Lamport's attempt to make TeX as easy to use as Brian Reid's Scribe(1), and indeed many/most of the concepts were "borrowed" directly from Scribe.

That being said, I'm pretty sure that the "--" and "---" substitutions are part of plain TeX, or, more probably, simply ligatures in the fonts used.

(1) Scribe was a lot less customizable than TeX, but it was far easier to use for 99% of the documents that were actually written in academic environments. It also had the great advantage that it didn't depend on a standardized set of fonts. You could use the same Scribe source and generate output for a Dover, a PostScript printer, a (paginated or non-paginated) ASCII file, a line printer, or virtually anything else just by specifying a command-line switch, and the compiler would read in a description of the device and learn what sorts of things it could do. (E.g., could it change the size of the font? Could it do boldface or underlined text? Could it do a half-line feed? Could it back up and overprint? Would that make things bold?) Scribe probably would have become the standard (picking up more of the TeX-like customizability) had it not been sold to a company that decided that they should charge for licenses, while TeX was free for the taking.

-- Evan Kirshenbaum +-- HP Laboratories |"You can't prove it isn't so!" is 1501 Page Mill Road, 1U, MS 1141 |as good as Q.E.D. in folk logic--as Palo Alto, CA 94304 |though it were necessary to submit |a piece of the moon to chemical (Email Removed) |analysis before you could be sure (650)857-7572 |that it was not made of green |cheese. http://www.kirshenbaum.net/ | Bergen Evans

Usenet · Answer

[nq:1]Simon R. Hughes (Email Removed) writes:[/nq]
[nq:2]Thus spake R F: LaTex is the interpreter so that ... the dashes. I am learning the code, which is TeX.[/nq]
[nq:1]LaTeX is (or at least was initially) nothing more nor less than a big set of TeX macros. The "latex" ... TeX as easy to use as Brian Reid's Scribe(1), and indeed many/most of the concepts were "borrowed" directly from Scribe.[/nq]
And LaTeX began by loading a slightly modified plain.tex
[nq:1]That being said, I'm pretty sure that the "--" and "---" substitutions are part of plain TeX, or, more probably, simply ligatures in the fonts used.[/nq]
Indeed, ligatures like ffi Below the plain level even, just a matter of fonts.
[nq:1](1) Scribe was a lot less customizable than TeX, but it was far easier to use for 99% of the ... been sold to a company that decided that they should charge for licenses, while TeX was free for the taking.[/nq]
And that made it out of date already in its time. Knuth's idea of writing output to a device independent .dvi file made more sense than adapting to the peculiarities (of by now long forgotten) hardware. It was the great portability and device independence that made TeX such a practical succes. I guess people had been willing to pay for it, if it hadn't been free. They also pay for some implementations.

Nowadays, with postscript and .pdf TeX has been adapted to write .pdf files directly.

Jan

Usenet · Answer

(Email Removed) (J. J. Lodder) writes:
[nq:2](1) Scribe was a lot less customizable than TeX, but ... charge for licenses, while TeX was free for the taking.[/nq]
[nq:1]And that made it out of date already in its time. Knuth's idea of writing output to a device independent ... by now long forgotten) hardware. It was the great portability and device independence that made TeX such a practical succes.[/nq]
Strong disagreement here. (And I speak as one who used both extensively for years and would probably have been considered a mid-level TeXpert at one time, having written a number of very low-level packages.) DVI made enormous assumptions about the capabilities of the resulting device. In particular, that it would be able to portray a bitmap, but also that it's paper was a particular size. The only thing really "device-independent" about it was that it didn't care about the resolution, and conversion to a printable representation was largely a simple matter of bitmap conversion. (Which, of course, required a conversion program for each device.) TeX even assumed its own set of bitmap fonts, at fixed sizes. The only thing that kept TeX from going away in the face of PostScript printers (besides the fanatical community) was that somebody hacked it to be able to use Adobe's fonts, but the resulting DVI files were usable only on PostScript printers.

In Scribe, on the other hand, the document source was the device-independent representation, and you only ran the compiler when you actually wanted output on your particular device. Not only did this allow a much wider range of devices, it also allowed the document to be reflowed to look good on the device being used.

The comparison that springs to mind is that TeX is PDF and Scribe was HTML. With TeX/PDF, you can get great formatting, and if you can render it at all it will always look the same. With Scribe/HTML, you have a much wider range of targets (Scribe->speech would be straightforward; DVI->speech would be a nightmare) and greater end-user control.

-- Evan Kirshenbaum +-- HP Laboratories |Its like grasping the difference 1501 Page Mill Road, 1U, MS 1141 |between what one usually considers Palo Alto, CA 94304 |a 'difficult' problem, and what |is* a difficult problem. The day (Email Removed) |one understands *why counting all (650)857-7572 |the molecules in the Universe isn't |difficult...there's the leap. http://www.kirshenbaum.net/ | Tina Marie Holmboe

Usenet · Answer

Thus spake J. J. Lodder:
[nq:1]Nowadays, with postscript and .pdf TeX has been adapted to write .pdf files directly.[/nq]
That's what I use it for. -- Simon R. Hughes

Usenet · Answer

[nq:1](Email Removed) (J. J. Lodder) writes:[/nq]
[nq:2] And that made it out of date already in ... and device independence that made TeX such a practical succes.[/nq]
[nq:1]Strong disagreement here. (And I speak as one who used both extensively for years and would probably have been considered ... In particular, that it would be able to portray a bitmap, but also that it's paper was a particular size.[/nq]
Not at all. It just says which character should by put where, in real coordinates. It doesn't say at all what the device driver and the device should do with that information. Thus I have seen TeX output rendered on a wide chain printer, with x-s representing pixels, at about three times the intended size. Usable, for proofing purposes. And also on some thermal kind of thing at a very fine resolution. It printed on 6" wide strips, iirc. A magnifying glass was needed to read the small print. All this when laserwriters were still undreamt of.
[nq:1]The only thing really "device-independent" about it was that it didn't care about the resolution, and conversion to a printable representation was largely a simple matter of bitmap conversion. (Which, of course, required a conversion program for each device.)[/nq]
Sure, that's the point.
[nq:1]TeX even assumed its own set of bitmap fonts, at fixed sizes. The only thing that kept TeX from going ... hacked it to be able to use Adobe's fonts, but the resulting DVI files were usable only on PostScript printers.[/nq]
Certainly not, it was for example quite possible to render TeX (including PS fonts) fairly well on an Apple ImageWriter, which was a needle device at 2*72 dpi, without any idea at all about Postscript. And no hacking was necessary to use arbitrary PostScript fonts. All that was needed was the import of the .afm files that came with the font. It may have been more difficult with other implementations though.
[nq:1]In Scribe, on the other hand, the document source was the device-independent representation, and you only ran the compiler when ... wider range of devices, it also allowed the document to be reflowed to look good on the device being used.[/nq]
Well, that is precisely what Knuth wanted to avoid at all cost. The good thing about TeX is that it will look the same, no matter how poorly rendered by some device, to all readers. You don't want your formulae reflowed (and probably garbled) by some device you know nothing about. Even when it looks good it may not be what you intended.
[nq:1]The comparison that springs to mind is that TeX is PDF and Scribe was HTML. With TeX/PDF, you can get ... have a much wider range of targets (Scribe->speech would be straightforward; DVI->speech would be a nightmare) and greater end-user control.[/nq]
Use dvi2ps, Postscript -> ASCII, speak ASCII should work,

Jan

Usenet · Answer

(Email Removed) (J. J. Lodder) writes:
[nq:2](Email Removed) (J. J. Lodder) writes: Strong disagreement here. (And ... bitmap, but also that it's paper was a particular size.[/nq]
[nq:1]Not at all. It just says which character should by put where, in real coordinates. It doesn't say at all what the device driver and the device should do with that information.[/nq]
But it assumes that the one processing the DVI file have access to (and can use) fonts that precisely match the metrics file. In practice that meant that for device independence, you had to use the same metafont-described bitmapped fonts that the author of the document used when he ran TeX. (I can't remember whether the bitmaps for characters used were actually stored in the DVI file or whether the machine that ran dvitowhatever had to have them around.)
[nq:1]Thus I have seen TeX output rendered on a wide chain printer, with x-s representing pixels, at about three times the intended size.[/nq]
Okay, "particular size up to a linear transformation". Have you ever seen TeX output rendered in a "usable" representation on such a printer? Or in an ASCII file? Could this exchange have easily been written in TeX in such a way that it would have given us the ASCII output we see here and nicely formatted output when printed out?
[nq:1]Usable, for proofing purposes. And also on some thermal kind of thing at a very fine resolution. It printed on 6" wide strips, iirc. A magnifying glass was needed to read the small print. All this when laserwriters were still undreamt of.[/nq]
Exactly. TeX was invented in the late '70s for publishing*, as in books and journals. Scribe was invented at the same time for *printing, as in letters, memos, manuals, handouts, dissertations, and tech reports. Both of them could take advantage of bitmap printers, which existed but were too expensive to assume people had, but only TeX required them.
[nq:2]TeX even assumed its own set of bitmap fonts, at ... the resulting DVI files were usable only on PostScript printers.[/nq]
[nq:1]Certainly not, it was for example quite possible to render TeX (including PS fonts) fairly well on an Apple ImageWriter, which was a needle device at 2*72 dpi, without any idea at all about Postscript.[/nq]
How exactly did that work? If TeX didn't know what the PostScript characters looked like (other than how big they were), and the ImageWriter (by your assertion) didn't know what they looked like, who told the printer where to put the dots?
[nq:1]And no hacking was necessary to use arbitrary PostScript fonts. All that was needed was the import of the .afm files that came with the font.[/nq]
And when you did that, you produced a "device independent" file that could only be printed on a device that had those fonts available. Not only wouldn't it print on the line printer, it wouldn't print on the LaserJet across the way, the Imagen down the hall, or the Dover in the next department.
[nq:1]It may have been more difficult with other implementations though.[/nq]
[nq:2]In Scribe, on the other hand, the document source was ... be reflowed to look good on the device being used.[/nq]
[nq:1]Well, that is precisely what Knuth wanted to avoid at all cost. The good thing about TeX is that it will look the same, no matter how poorly rendered by some device, to all readers.[/nq]
For publishing, sure. For most writing, not at all. As I'm typing this, I really don't care what font you're going to read it in, what margins you choose, or how your reader will cope with this. I do care to some extent that it will recognize that this is a paragraph of text, and that the prior paragraph is an excerpt from the message I'm replying to.
[nq:1]You don't want your formulae reflowed (and probably garbled) by some device you know nothing about. Even when it looks good it may not be what you intended.[/nq]
Ninety-nine percent of documents written contain no formulas that could be hurt by such things. Especially if the formula itself is specified structurally. But I'll grant you that TeX was better than Scribe for documents that contained a lot of mathematical formulas.
[nq:2]The comparison that springs to mind is that TeX is ... straightforward; DVI->speech would be a nightmare) and greater end-user control.[/nq]
[nq:1]Use dvi2ps, Postscript -> ASCII, speak ASCII should work,[/nq]
Unless things have changed a lot, this would be horrible. First of all, the versions of dvi2ps I'm familiar with produced PostScript bitmaps (which is one of the reasons DVI files printed so slowly) unless the TeX input specified using PostScript fonts. Secondly, DVI files don't necessarily contain all the characters in the logical order so what you'd get would probably be a garbled mess. Third, the resulting files, even for PostScript fonts, make no distinction between spaces inserted between words, those inserted between sentences, and those inserted just for kerning, so it would be hard to find word and sentence breaks. Fourth, all DVI files have is font changes. There's no way to tell if, say, a shift from italics to roman is meant to signify the end of emphasis or nested emphasis, a distinction preserved in the Scribe source (and LaTeX, though not plain TeX, which lacked "\em"). Fifth, if you were formatting a document for speech, you wouldn't insert things like page numbers, headers, and footers, and you would probably handle footnotes differently.

-- Evan Kirshenbaum +-- HP Laboratories |When you rewrite a compiler from 1501 Page Mill Road, 1U, MS 1141 |scratch, you sometimes fix things Palo Alto, CA 94304 |you didn't know were broken. | Larry Wall (Email Removed) (650)857-7572

http://www.kirshenbaum.net/

Usenet · Answer

[nq:1](Email Removed) (J. J. Lodder) writes:[/nq]
[nq:2] Not at all. It just says which character should ... device driver and the device should do with that information.[/nq]
[nq:1]But it assumes that the one processing the DVI file have access to (and can use) fonts that precisely match the metrics file.[/nq]
No, you can put any character of any dimension you happen to have at the specified position. The result may of course differ from what was intended. Printing a document on a fixed pitch line printer using only x-s certainly won't match the font metrics; yet the result is usable for some puposes.
[nq:1]In practice that meant that for device independence, you had to use the same metafont-described bitmapped fonts that the author ... used were actually stored in the DVI file or whether the machine that ran dvitowhatever had to have them around.)[/nq]
The dvi driver looked them up, iirc. You needed a screen version and a print version of each font. (lots of floppies, with 72 pt pixel files, one for each font)

But that was only in the very beginning. Postcript versions of the fonts were made when postscript became available. (And MetaFont dropped into (near) oblivion.)
[nq:2]Thus I have seen TeX output rendered on a wide chain printer, with x-s representing pixels, at about three times the intended size.[/nq]
[nq:1]Okay, "particular size up to a linear transformation". Have you ever seen TeX output rendered in a "usable" representation on such a printer?[/nq]
Yes, it was quite useable. You looked at it from a distance. The only drawback was the paper consumption, and having to carry it all.
[nq:1]Or in an ASCII file? Could this exchange have easily been written in TeX in such a way that it would have given us the ASCII output we see here and nicely formatted output when printed out?[/nq]
Why not? Feeding this dialog to TeX, with a suitable macro package could pretty-print it. It could even be made to set my text to \raggedright near 50 chars, yours to blockwrapped to 80, by counting the quote chars, if we don't snip to much in between.
[nq:2]Usable, for proofing purposes. And also on some thermal kind ... small print. All this when laserwriters were still undreamt of.[/nq]
[nq:1]Exactly. TeX was invented in the late '70s for publishing, as in books and journals. Scribe was invented at the ... take advantage of bitmap printers, which existed but were too expensive to assume people had, but only TeX required them.[/nq]
No comment, don't remember enough about Scribe. And no, TeX was (and is) quite suiable for writing letters, memos etc, even before LaTeX. It just took some macros.
[nq:2]Certainly not, it was for example quite possible to render ... at *** dpi, without any idea at all about Postscript.[/nq]
[nq:1]How exactly did that work?[/nq]
I'l have to search, my good old MacPlus is long since dead.
[nq:1]If TeX didn't know what the PostScript characters looked like (other than how big they were), and the ImageWriter (by your assertion) didn't know what they looked like, who told the printer where to put the dots?[/nq]
Sorry, took some time, but I found the back-ups from that era to be no longer usable. There were on floppies. Adobe provided a rasterizer, iirc. I am sure I did use third party (shareware) postscript fonts with TeX, at the time.
[nq:2]And no hacking was necessary to use arbitrary PostScript fonts. ... import of the .afm files that came with the font.[/nq]
[nq:1]And when you did that, you produced a "device independent" file that could only be printed on a device that ... wouldn't print on the LaserJet across the way, the Imagen down the hall, or the Dover in the next department.[/nq]
Sure, 'device independent' doesn't mean font independent. If you used 'strange' fonts you could run dvi2ps for printing elsewhere, and hope for the best. dvi files weren't intended to be portable anyway. You sent the TeX source, and the additional postscript fonts, if any, and let the other side handle it their way.
[nq:2]It may have been more difficult with other implementations though. ... matter how poorly rendered by some device, to all readers.[/nq]
[nq:1]For publishing, sure. For most writing, not at all. As I'm typing this, I really don't care what font you're ... this is a paragraph of text, and that the prior paragraph is an excerpt from the message I'm replying to.[/nq]
Usenet is (and should be) plain text only.
[nq:2]You don't want your formulae reflowed (and probably garbled) by ... it looks good it may not be what you intended.[/nq]
[nq:1]Ninety-nine percent of documents written contain no formulas that could be hurt by such things. Especially if the formula itself is specified structurally. But I'll grant you that TeX was better than Scribe for documents that contained a lot of mathematical formulas.[/nq]
And better than Quark for formatting newsprint. Even its hyphenation is still better than what my (supposedly quality) newspaper gives me.
[nq:2]Use dvi2ps, Postscript -> ASCII, speak ASCII should work,[/nq]
[nq:1]Unless things have changed a lot, this would be horrible. First of all, the versions of dvi2ps I'm familiar with produced PostScript bitmaps (which is one of the reasons DVI files printed so slowly) unless the TeX input specified using PostScript fonts.[/nq]
That is long ago. Only the very first TeX implementation I saw had bitmapped fonts. I never succeeded in getting it to print. All later ones used postscript fonts. Your impression may be due to running inadequate implementions of TeX, under DOS or Windoze?
[nq:1]files don't necessarily contain all the characters in the logical order so what you'd get would probably be a garbled mess.[/nq]
Not necessaily, but in practice you can often read large blocks of text straight from the postscript file.
[nq:1]Third, the resulting files, even for PostScript fonts, make no distinction between spaces inserted between words, those inserted between sentences, and those inserted just for kerning, so it would be hard to find word and sentence breaks.[/nq]
It is not that bad. I know, for I did have to extract plain text from TeX for someone who wanted to recycle it as (shudder) MSWord.
[nq:1]Fourth, all DVI files have is font changes. There's no way to tell if, say, a shift from italics to ... document for speech, you wouldn't insert things like page numbers, headers, and footers, and you would probably handle footnotes differently.[/nq]
Sure, it would be better (but much more work) to write some TeX macros for speakable output.

Best,

Jan

En and Em dashes

Top answer

Get the Weekly English Kit 📬

Related Questions