Re: More indexing bogosity
Steve Litt wrote:
This time, when inserting see and seealso records, I'm forced to use ERT
for each of the curly braces or else part of the seealso text doesn't
show up and an erroneous page number is shown. This is a huge time sink.
The big change in 1.6.x indexing is that indexes now behave as normal text. So
you have to use ERT whereever you would need ERT in normal text. Since |see{}
is a LaTeX macro (just using | instead of \), ERT is not bogus at all
here.
The gain of this change is that you are not forced to write \ubergro\ss but
can simply write übergroß in all encodings. And you can directly emphasize
entries, without the need to write \emph{my entry}.
Jürgen
Re: More indexing bogosity
On Thursday 10 September 2009 04:14:44 Jürgen Spitzmüller wrote:
Steve Litt wrote:
This time, when inserting see and seealso records, I'm forced to use ERT
for each of the curly braces or else part of the seealso text doesn't
show up and an erroneous page number is shown. This is a huge time sink.
The big change in 1.6.x indexing is that indexes now behave as normal text.
So you have to use ERT whereever you would need ERT in normal text. Since
|see{} is a LaTeX macro (just using | instead of \), ERT is not bogus
at all here.
Let me refine the concept of bogus. True, the ERT isn't bogus from a data
consistency point of view, nor is it bogus from a logical point of view. But
it's very bogus from the point of view of the user's workflow. The user must
now, twice per index see or seealso entry, press Ctrl+L and type in a curly
brace. If the user doesn't use T1 output character set, he has to do something
similar with every regular index entry with a range. This significantly slows
the user, makes the user think about something other than the content, and can
cause temporary failure to an unsophisticated user.
The point of computer programming is to make things faster and easier for the
user. Otherwise, why not just use TeX? This change makes things slower and
harder while indexing.
The gain of this change is that you are not forced to write \ubergro\ss
but can simply write übergroß in all encodings. And you can directly
emphasize entries, without the need to write \emph{my entry}.
That's definitely a gain. The trouble is, it comes at the cost of a loss for
ranges and sees and seealsos. It also breaks formerly working documents. I'm
sure there are ways to implement the advantages without the disadvantages.
Here are a few:
* Use a dialog box to define each element of an index entry.
* Have a single character to escape |, { and }
* Have an option to evaluate indexes the way 1.5.x did
If nothing is done about this, it won't harm me. I write only English so T1 is
fine with me. I find it better to put see and seealso entries in a separate
file
so I won't have to ERT twice per entry. Even if things change some more so T1
doesn't work for me, I already have a kludge script to correct the LyX
produced LaTeX. But these opportunities aren't available to a lot of users,
and they will be slowed by this new method of parsing LyX index entries.
StevET
Steve Litt
Recession Relief Package
http://www.recession-relief.US
Twitter: http://www.twitter.com/stevelitt
Re: More indexing bogosity
Steve Litt wrote:
Let me refine the concept of bogus. True, the ERT isn't bogus from a data
consistency point of view, nor is it bogus from a logical point of view.
But it's very bogus from the point of view of the user's workflow. The
user must now, twice per index see or seealso entry, press Ctrl+L and type
in a curly brace. If the user doesn't use T1 output character set, he has
to do something similar with every regular index entry with a range. This
significantly slows the user, makes the user think about something other
than the content, and can cause temporary failure to an unsophisticated
user.
I'd consider everything behind | in an index entry sophisticated, because
it requires LaTeX knowledge. I'd say that unsophisticated users would rather
expect { to be typeset as { rather than to be verbatim LaTeX.
In LyX, all verbatim LaTeX has to be embraced by ERT. Why should we make an
exception just for indexes? I'd consider that conceptually bogus.
The point of computer programming is to make things faster and easier for
the user. Otherwise, why not just use TeX? This change makes things slower
and harder while indexing.
It makes some things harder while making others faster. We considered the gain
bigger than the burden.
The gain of this change is that you are not forced to write \ubergro\ss
but can simply write übergroß in all encodings. And you can directly
emphasize entries, without the need to write \emph{my entry}.
That's definitely a gain. The trouble is, it comes at the cost of a loss
for ranges and sees and seealsos. It also breaks formerly working
documents.
It's not supposed to break formerly working documents. In other words: the
braces should be wrapped in ERT by the lyx2lyx converter. I just verified that
this works. If it doesn't for you, please provide a test case.
I know that | need to be wrapped as well, and they aren't yet. But you know
that I have already developed a patch to fix this bug.
I'm sure there are ways to implement the advantages without the
disadvantages. Here are a few:
* Use a dialog box to define each element of an index entry.
* Have a single character to escape |, { and }
* Have an option to evaluate indexes the way 1.5.x did
As already stated, the way would be to support |see and the like natively.
Unless we do that, it is treated as any other verbatim LaTeX.
However, you can easily restore the 1.5.x behaviour:
copy the file stdinsets.onc from the System directory to the user directory,
and in the entry of InsetLayout Index, add the line
PassThru true
Then all content of the inset will be treated verbatim (as in ERT). Of course,
you will lose all the 1.6.x index improvements with this change.
If nothing is done about this, it won't harm me. I write only English so T1
is fine with me. I find it better to put see and seealso entries in a
separate file so I won't have to ERT twice per entry. Even if things
change some more so T1 doesn't work for me, I already have a kludge script
to correct the LyX produced LaTeX. But these opportunities aren't
available to a lot of users, and they will be slowed by this new method of
parsing LyX index entries.
As said, I have proposed a solution for the T1 problem. So no kludge script is
needed.
StevET
Steve Litt
Recession Relief Package
http://www.recession-relief.US
Twitter: http://www.twitter.com/stevelitt
Jürgen
Re: More indexing bogosity
Steve Litt wrote:
This time, when inserting see and seealso records, I'm forced to use ERT
for each of the curly braces or else part of the seealso text doesn't
show up and an erroneous page number is shown. This is a huge time sink.
The big change in 1.6.x indexing is that indexes now behave as normal text. So
you have to use ERT whereever you would need ERT in normal text. Since |see{}
is a LaTeX macro (just using | instead of \), ERT is not bogus at all
here.
The gain of this change is that you are not forced to write \ubergro\ss but
can simply write übergroß in all encodings. And you can directly emphasize
entries, without the need to write \emph{my entry}.
Jürgen
Re: More indexing bogosity
On Thursday 10 September 2009 04:14:44 Jürgen Spitzmüller wrote:
Steve Litt wrote:
This time, when inserting see and seealso records, I'm forced to use ERT
for each of the curly braces or else part of the seealso text doesn't
show up and an erroneous page number is shown. This is a huge time sink.
The big change in 1.6.x indexing is that indexes now behave as normal text.
So you have to use ERT whereever you would need ERT in normal text. Since
|see{} is a LaTeX macro (just using | instead of \), ERT is not bogus
at all here.
Let me refine the concept of bogus. True, the ERT isn't bogus from a data
consistency point of view, nor is it bogus from a logical point of view. But
it's very bogus from the point of view of the user's workflow. The user must
now, twice per index see or seealso entry, press Ctrl+L and type in a curly
brace. If the user doesn't use T1 output character set, he has to do something
similar with every regular index entry with a range. This significantly slows
the user, makes the user think about something other than the content, and can
cause temporary failure to an unsophisticated user.
The point of computer programming is to make things faster and easier for the
user. Otherwise, why not just use TeX? This change makes things slower and
harder while indexing.
The gain of this change is that you are not forced to write \ubergro\ss
but can simply write übergroß in all encodings. And you can directly
emphasize entries, without the need to write \emph{my entry}.
That's definitely a gain. The trouble is, it comes at the cost of a loss for
ranges and sees and seealsos. It also breaks formerly working documents. I'm
sure there are ways to implement the advantages without the disadvantages.
Here are a few:
* Use a dialog box to define each element of an index entry.
* Have a single character to escape |, { and }
* Have an option to evaluate indexes the way 1.5.x did
If nothing is done about this, it won't harm me. I write only English so T1 is
fine with me. I find it better to put see and seealso entries in a separate
file
so I won't have to ERT twice per entry. Even if things change some more so T1
doesn't work for me, I already have a kludge script to correct the LyX
produced LaTeX. But these opportunities aren't available to a lot of users,
and they will be slowed by this new method of parsing LyX index entries.
StevET
Steve Litt
Recession Relief Package
http://www.recession-relief.US
Twitter: http://www.twitter.com/stevelitt
Re: More indexing bogosity
Steve Litt wrote:
Let me refine the concept of bogus. True, the ERT isn't bogus from a data
consistency point of view, nor is it bogus from a logical point of view.
But it's very bogus from the point of view of the user's workflow. The
user must now, twice per index see or seealso entry, press Ctrl+L and type
in a curly brace. If the user doesn't use T1 output character set, he has
to do something similar with every regular index entry with a range. This
significantly slows the user, makes the user think about something other
than the content, and can cause temporary failure to an unsophisticated
user.
I'd consider everything behind | in an index entry sophisticated, because
it requires LaTeX knowledge. I'd say that unsophisticated users would rather
expect { to be typeset as { rather than to be verbatim LaTeX.
In LyX, all verbatim LaTeX has to be embraced by ERT. Why should we make an
exception just for indexes? I'd consider that conceptually bogus.
The point of computer programming is to make things faster and easier for
the user. Otherwise, why not just use TeX? This change makes things slower
and harder while indexing.
It makes some things harder while making others faster. We considered the gain
bigger than the burden.
The gain of this change is that you are not forced to write \ubergro\ss
but can simply write übergroß in all encodings. And you can directly
emphasize entries, without the need to write \emph{my entry}.
That's definitely a gain. The trouble is, it comes at the cost of a loss
for ranges and sees and seealsos. It also breaks formerly working
documents.
It's not supposed to break formerly working documents. In other words: the
braces should be wrapped in ERT by the lyx2lyx converter. I just verified that
this works. If it doesn't for you, please provide a test case.
I know that | need to be wrapped as well, and they aren't yet. But you know
that I have already developed a patch to fix this bug.
I'm sure there are ways to implement the advantages without the
disadvantages. Here are a few:
* Use a dialog box to define each element of an index entry.
* Have a single character to escape |, { and }
* Have an option to evaluate indexes the way 1.5.x did
As already stated, the way would be to support |see and the like natively.
Unless we do that, it is treated as any other verbatim LaTeX.
However, you can easily restore the 1.5.x behaviour:
copy the file stdinsets.onc from the System directory to the user directory,
and in the entry of InsetLayout Index, add the line
PassThru true
Then all content of the inset will be treated verbatim (as in ERT). Of course,
you will lose all the 1.6.x index improvements with this change.
If nothing is done about this, it won't harm me. I write only English so T1
is fine with me. I find it better to put see and seealso entries in a
separate file so I won't have to ERT twice per entry. Even if things
change some more so T1 doesn't work for me, I already have a kludge script
to correct the LyX produced LaTeX. But these opportunities aren't
available to a lot of users, and they will be slowed by this new method of
parsing LyX index entries.
As said, I have proposed a solution for the T1 problem. So no kludge script is
needed.
StevET
Steve Litt
Recession Relief Package
http://www.recession-relief.US
Twitter: http://www.twitter.com/stevelitt
Jürgen
Re: More indexing bogosity
Steve Litt wrote:
> This time, when inserting see and seealso records, I'm forced to use ERT
> for each of the curly braces or else part of the seealso text doesn't
> show up and an erroneous page number is shown. This is a huge time sink.
The big change in 1.6.x indexing is that indexes now behave as normal text. So
you have to use ERT whereever you would need ERT in normal text. Since |see{}
is a LaTeX macro (just using "|" instead of "\"), ERT is not bogus at all
here.
The gain of this change is that you are not forced to write \"ubergro\ss but
can simply write übergroß in all encodings. And you can directly emphasize
entries, without the need to write \emph{my entry}.
Jürgen
Re: More indexing bogosity
On Thursday 10 September 2009 04:14:44 Jürgen Spitzmüller wrote:
> Steve Litt wrote:
> > This time, when inserting see and seealso records, I'm forced to use ERT
> > for each of the curly braces or else part of the seealso text doesn't
> > show up and an erroneous page number is shown. This is a huge time sink.
>
> The big change in 1.6.x indexing is that indexes now behave as normal text.
> So you have to use ERT whereever you would need ERT in normal text. Since
> |see{} is a LaTeX macro (just using "|" instead of "\"), ERT is not bogus
> at all here.
Let me refine the concept of "bogus". True, the ERT isn't bogus from a data
consistency point of view, nor is it bogus from a logical point of view. But
it's very bogus from the point of view of the user's workflow. The user must
now, twice per index see or seealso entry, press Ctrl+L and type in a curly
brace. If the user doesn't use T1 output character set, he has to do something
similar with every regular index entry with a range. This significantly slows
the user, makes the user think about something other than the content, and can
cause temporary failure to an unsophisticated user.
The point of computer programming is to make things faster and easier for the
user. Otherwise, why not just use TeX? This change makes things slower and
harder while indexing.
>
> The gain of this change is that you are not forced to write \"ubergro\ss
> but can simply write übergroß in all encodings. And you can directly
> emphasize entries, without the need to write \emph{my entry}.
That's definitely a gain. The trouble is, it comes at the cost of a loss for
ranges and sees and seealsos. It also breaks formerly working documents. I'm
sure there are ways to implement the advantages without the disadvantages.
Here are a few:
* Use a dialog box to define each element of an index entry.
* Have a single character to escape |, { and }
* Have an option to evaluate indexes the way 1.5.x did
If nothing is done about this, it won't harm me. I write only English so T1 is
fine with me. I find it better to put see and seealso entries in a separate
file
so I won't have to ERT twice per entry. Even if things change some more so T1
doesn't work for me, I already have a kludge script to correct the LyX
produced LaTeX. But these opportunities aren't available to a lot of users,
and they will be slowed by this new method of parsing LyX index entries.
StevET
Steve Litt
Recession Relief Package
http://www.recession-relief.US
Twitter: http://www.twitter.com/stevelitt
Re: More indexing bogosity
Steve Litt wrote:
> Let me refine the concept of "bogus". True, the ERT isn't bogus from a data
> consistency point of view, nor is it bogus from a logical point of view.
> But it's very bogus from the point of view of the user's workflow. The
> user must now, twice per index see or seealso entry, press Ctrl+L and type
> in a curly brace. If the user doesn't use T1 output character set, he has
> to do something similar with every regular index entry with a range. This
> significantly slows the user, makes the user think about something other
> than the content, and can cause temporary failure to an unsophisticated
> user.
I'd consider everything behind "|" in an index entry "sophisticated", because
it requires LaTeX knowledge. I'd say that unsophisticated users would rather
expect "{" to be typeset as "{" rather than to be verbatim LaTeX.
In LyX, all verbatim LaTeX has to be embraced by ERT. Why should we make an
exception just for indexes? I'd consider that conceptually bogus.
> The point of computer programming is to make things faster and easier for
> the user. Otherwise, why not just use TeX? This change makes things slower
> and harder while indexing.
It makes some things harder while making others faster. We considered the gain
bigger than the burden.
> > The gain of this change is that you are not forced to write \"ubergro\ss
> > but can simply write übergroß in all encodings. And you can directly
> > emphasize entries, without the need to write \emph{my entry}.
>
> That's definitely a gain. The trouble is, it comes at the cost of a loss
> for ranges and sees and seealsos. It also breaks formerly working
> documents.
It's not supposed to break formerly working documents. In other words: the
braces should be wrapped in ERT by the lyx2lyx converter. I just verified that
this works. If it doesn't for you, please provide a test case.
I know that "|" need to be wrapped as well, and they aren't yet. But you know
that I have already developed a patch to fix this bug.
> I'm sure there are ways to implement the advantages without the
> disadvantages. Here are a few:
>
> * Use a dialog box to define each element of an index entry.
> * Have a single character to escape |, { and }
> * Have an option to evaluate indexes the way 1.5.x did
As already stated, the way would be to support "|see" and the like natively.
Unless we do that, it is treated as any other verbatim LaTeX.
However, you can easily restore the 1.5.x behaviour:
copy the file stdinsets.onc from the System directory to the user directory,
and in the entry of InsetLayout Index, add the line
PassThru true
Then all content of the inset will be treated verbatim (as in ERT). Of course,
you will lose all the 1.6.x index improvements with this change.
> If nothing is done about this, it won't harm me. I write only English so T1
> is fine with me. I find it better to put see and seealso entries in a
> separate file so I won't have to ERT twice per entry. Even if things
> change some more so T1 doesn't work for me, I already have a kludge script
> to correct the LyX produced LaTeX. But these opportunities aren't
> available to a lot of users, and they will be slowed by this new method of
> parsing LyX index entries.
As said, I have proposed a solution for the T1 problem. So no kludge script is
needed.
> StevET
>
> Steve Litt
> Recession Relief Package
> http://www.recession-relief.US
> Twitter: http://www.twitter.com/stevelitt
>
Jürgen
Re: More indexing bogosity WORKED AROUND
On Wednesday 09 September 2009 15:05:17 Steve Litt wrote:
Hi all,
I'm deep in the middle of indexing my new book and found yet another
problem, and this one can't be worked around by setting
Tools-preferences-output-
latex-Tex_encoding to T1.
This time, when inserting see and seealso records, I'm forced to use ERT
for each of the curly braces or else part of the seealso text doesn't show
up and an erroneous page number is shown. This is a huge time sink.
As a workaround, is there a way I can have all my see and seealso entries
in an external file and import it? I seem to remember that can be done, and
hopefully it would work around this time sink.
Thanks
SteveT
OK, I found it. You can put all see and seealso index records in a separate
file. For instance, I put it in ./seealso.inc. So ./seealso.inc has lots
records like this:
\index{Joneses|see{keeping up with the Joneses}}
Then, in the document preamble, I place the following line:
\input{./seealso.inc}
That does the trick, and I don't have to do any ERT maneuvers.
SteveT
Steve Litt
Recession Relief Package
http://www.recession-relief.US
Twitter: http://www.twitter.com/stevelitt
Re: More indexing bogosity WORKED AROUND
On Wednesday 09 September 2009 15:05:17 Steve Litt wrote:
Hi all,
I'm deep in the middle of indexing my new book and found yet another
problem, and this one can't be worked around by setting
Tools-preferences-output-
latex-Tex_encoding to T1.
This time, when inserting see and seealso records, I'm forced to use ERT
for each of the curly braces or else part of the seealso text doesn't show
up and an erroneous page number is shown. This is a huge time sink.
As a workaround, is there a way I can have all my see and seealso entries
in an external file and import it? I seem to remember that can be done, and
hopefully it would work around this time sink.
Thanks
SteveT
OK, I found it. You can put all see and seealso index records in a separate
file. For instance, I put it in ./seealso.inc. So ./seealso.inc has lots
records like this:
\index{Joneses|see{keeping up with the Joneses}}
Then, in the document preamble, I place the following line:
\input{./seealso.inc}
That does the trick, and I don't have to do any ERT maneuvers.
SteveT
Steve Litt
Recession Relief Package
http://www.recession-relief.US
Twitter: http://www.twitter.com/stevelitt
Re: More indexing bogosity
On Wednesday 09 September 2009 15:05:17 Steve Litt wrote:
> Hi all,
>
> I'm deep in the middle of indexing my new book and found yet another
> problem, and this one can't be worked around by setting
> Tools->preferences->output-
>
> >latex->Tex_encoding to T1.
>
> This time, when inserting see and seealso records, I'm forced to use ERT
> for each of the curly braces or else part of the seealso text doesn't show
> up and an erroneous page number is shown. This is a huge time sink.
>
> As a workaround, is there a way I can have all my see and seealso entries
> in an external file and import it? I seem to remember that can be done, and
> hopefully it would work around this time sink.
>
> Thanks
>
> SteveT
OK, I found it. You can put all see and seealso index records in a separate
file. For instance, I put it in ./seealso.inc. So ./seealso.inc has lots
records like this:
\index{Joneses|see{keeping up with the Joneses}}
Then, in the document preamble, I place the following line:
\input{./seealso.inc}
That does the trick, and I don't have to do any ERT maneuvers.
SteveT
Steve Litt
Recession Relief Package
http://www.recession-relief.US
Twitter: http://www.twitter.com/stevelitt
