openLilyLib website
Hi all, as a starting point for a - hopefully - comprehensive documentation effort I have finally updated https://openlilylib.org with a completely new website, which I'd like to have some feedback about and contributions for. There are several parts to that effort, most of which are essentially not started yet. * A general introduction website. This is basically complete and should finally give a proper introduction about what OLL "is" and how it can be made to work * Independent sub-sites for each OLL package. These have not been written at all, only the links to empty starting pages work without 404 errors. * I've settled with MkDocs (https://www.mkdocs.org), which seems to provide what I need, especially a suitable way to hook into and extend to our needs. * Each sub-site is maintained in a separate Git repository and included as a Git submodule, so it should be straightforward to manage independent authoring of the documentation by the respective package maintainers. * There's a link to a contributor's guide, which is also essentially empty, except for an entry page. What I have so far is an infrastructure for textual, Markdown-authored manuals, although I have already created a plugin for LilyPond syntax highlighting using python-ly ( https://github.com/uliska/markdown-lilypond/). What I really *want* to have but have no idea so far how to achieve is additional code/API documentation retrieved from the actual source files. There should be a tool to retrieve that from comments (or actual signatures?), resulting in either HTML or Markdown documentation that can be automatically integrated in the "manual-style" documentation. Best Urs
Re: openLilyLib website
Dear Urs, it’s great that you found time for that. It looks nice and I think MkDocs is a good choice (would have been my first choice, too). While speaking about documentation, I very much like this approach: https://www.divio.com/blog/documentation/ Some nitpicks from me: - Should everyone see the "Edit this page" pencils? - Those links are broken for me - If you descend into a package documentation, "Home" will not bring you to openlilylib.org but to the page of the package. While that is probably a consequence of your fourth bullet point, I find it confusing. - Do you like the brown color? > * Each sub-site is maintained in a separate Git repository and >included as a Git submodule, so it should be straightforward to >manage independent authoring of the documentation by the respective >package maintainers. While this theoretically makes sense to separate the packages' documentation, I would run from git submodule. But perhaps your use case is easier than where my experience comes from and you can still change that decision later in case it makes things easier. > What I really *want* to have but have no idea so far how to achieve is > additional code/API documentation retrieved from the actual source > files. There should be a tool to retrieve that from comments (or actual > signatures?), What do you mean by signatures? IMHO, this makes sense for "reference-type" documentation in addition to other parts of the documentation (cf. link at the top). Cheers, Joram
Re: openLilyLib website
Hi Joram, thanks for the ffecback. Am Donnerstag, den 20.02.2020, 23:33 +0100 schrieb Noeck: > Dear Urs, > > it’s great that you found time for that. It looks nice and I think > MkDocs is a good choice (would have been my first choice, too). > > While speaking about documentation, I very much like this approach: > https://www.divio.com/blog/documentation/ > I'll definitely have a look, but only after replying ... > Some nitpicks from me: > > - Should everyone see the "Edit this page" pencils? Yes. Everybody should be encouraged to join this project ... >- Those links are broken for me ... which is of course only possible if the links aren't broken. ( https://github.com/openlilylib-documentation/main-site/issues/1) > - If you descend into a package documentation, "Home" will not bring > you > to openlilylib.org but to the page of the package. While that is > probably a consequence of your fourth bullet point, I find it > confusing. I'm not fully clear how to solve the underlying issue. In fact "Home" brings you the the home page of the package while "Main Book" brings you to the top of the whole site. I agree it may be confusing. The thing is, this site is *not* an actual MkDocs site but a nested set of MkDocs sites. This brings some additional potential for more complex set-ups and separation of related sub-books, but it comes at a cost with complexity. > - Do you like the brown color? I haven't given much thought on the colors yet. Each sub-book should choose its individual color, and the theme has only a limited number of preconfigured colors ( https://squidfunk.github.io/mkdocs-material/getting-started/#color-palette ). At one point I think we'll want to create a fork of the theme and add more colors, maybe even with some aesthetic specificity. But that's nothing that has to be decided at this point. (And I don't dislike that brown ...). > > > * Each sub-site is maintained in a separate Git repository and > >included as a Git submodule, so it should be straightforward to > >manage independent authoring of the documentation by the > > respective > >package maintainers. > > While this theoretically makes sense to separate the packages' > documentation, I would run from git submodule. But perhaps your use > case > is easier than where my experience comes from and you can still > change > that decision later in case it makes things easier. I don't have any significant experience with submodules. But while I was always a bit worried about them I think this use case is pretty safe because we're not talking about libraries some code depends on but essentially they're independent chunks of content. If a submodule is out of date or whatever simply that part of the site will look bad or out of date. > > > What I really *want* to have but have no idea so far how to achieve > > is > > additional code/API documentation retrieved from the actual source > > files. There should be a tool to retrieve that from comments (or > > actual > > signatures?), > > What do you mean by signatures? AFAICT there are basically two types of source file documentation systems: retrieving explicit documentation comments (aka docstrings) or analyzing the defined classes and functions (i.e. their signatures). Or a combination where the documentation software analyzes the function signatures but takes comments for their explanations. > IMHO, this makes sense for "reference-type" documentation in addition > to > other parts of the documentation (cf. link at the top). As said I'll read up on this, but "reference-type" "in addition" sounds pretty much exactly like what I have in mind. Best Urs > > Cheers, > Joram >
Re: openLilyLib website
On 02/20, Urs Liska wrote: > What I really *want* to have but have no idea so far how to achieve is > additional code/API documentation retrieved from the actual source > files. There should be a tool to retrieve that from comments (or actual > signatures?), resulting in either HTML or Markdown documentation that > can be automatically integrated in the "manual-style" documentation. Do you want something similar [Sphinx][1] or [Doxygen][2]? Are the source files in Scheme, Lilypond, or a combination of both? [1]: https://en.wikipedia.org/wiki/Sphinx_(documentation_generator) [2]: https://en.wikipedia.org/wiki/Doxygen signature.asc Description: PGP signature
Re: openLilyLib website
Am Freitag, den 21.02.2020, 18:52 -0800 schrieb ma...@masonhock.com: > On 02/20, Urs Liska wrote: > > What I really *want* to have but have no idea so far how to achieve > > is > > additional code/API documentation retrieved from the actual source > > files. There should be a tool to retrieve that from comments (or > > actual > > signatures?), resulting in either HTML or Markdown documentation > > that > > can be automatically integrated in the "manual-style" > > documentation. > > Do you want something similar [Sphinx][1] or [Doxygen][2]? Yes, something similar, but IISC these tools (which I had of course taken notice of earlier) are limited to their set of languages. Parsing documentation comments or docstrings and producing API reference is what is still needed. One aspect that might be different from "traditional" API reference documentation in a case like openLilyLib (also like LaTeX packages) is that there are *end-user*-facing commands and *internal* commands. Internal commands (i.e. auxiliary functions) may or may not be interesting for users outside the package development, but probably the documentation should be kept separate in some way. For example by providing totally separate pages/trees or listing the end-user functionality first on a page, followed by the internal stuff. In addition I think the end-user functionality whould be documented on a package-based page while the internal API should be documented per source file. But I haven't fully thought through that topic yet. Craig Dabelstein did some research and suggested NaturalDocs ( https://www.naturaldocs.org/) which looks nice and suitably extensible/configurable for our purpose. I'm a little bit wary about the fact that it's a project with a single contributor, and I know what I'm talking about ... (also I *think* it would be good to have a Python-based tool because it might be easier to extend and to integrate with the other stuff. > Are the > source files in Scheme, Lilypond, or a combination of both? Of course both. The end-user interface is usually LilyPond, but the functionality is all LilyPond, Scheme-in-LilyPond, and Scheme-in-Scheme-files. Urs > > [1]: https://en.wikipedia.org/wiki/Sphinx_(documentation_generator) > > [2]: https://en.wikipedia.org/wiki/Doxygen
Re: openLilyLib website
Hi Urs, this looks great! As often, I can’t offer much more than applause and some nitpicking… On the ‘About LilyPond’ I’d rather write “GNU LilyPond is a music notation program /strongly inspired by/ traditional craftsmanship”. After all, the process of working with LilyPond doesn’t resemble traditional craftsmanship at all… The use of the word ‘domain’ in the About/LilyPond and /openLilyLib seems quite technical considering IIUC the site is supposed to be read also by people without a background in software development… ‘Get started/Install openLilyLib’ has an instance of [oll-core} that seems unintentional. s/documenation/documentation … but generally, the texts seem very appropriate and useful :-) Very best regards Simon On 20.02.20 22:44, Urs Liska wrote: Hi all, as a starting point for a - hopefully - comprehensive documentation effort I have finally updated https://openlilylib.org with a completely new website, which I'd like to have some feedback about and contributions for. There are several parts to that effort, most of which are essentially not started yet. * A general introduction website. This is basically complete and should finally give a proper introduction about what OLL "is" and how it can be made to work * Independent sub-sites for each OLL package. These have not been written at all, only the links to empty starting pages work without 404 errors. * I've settled with MkDocs (https://www.mkdocs.org), which seems to provide what I need, especially a suitable way to hook into and extend to our needs. * Each sub-site is maintained in a separate Git repository and included as a Git submodule, so it should be straightforward to manage independent authoring of the documentation by the respective package maintainers. * There's a link to a contributor's guide, which is also essentially empty, except for an entry page. What I have so far is an infrastructure for textual, Markdown-authored manuals, although I have already created a plugin for LilyPond syntax highlighting using python-ly ( https://github.com/uliska/markdown-lilypond/). What I really *want* to have but have no idea so far how to achieve is additional code/API documentation retrieved from the actual source files. There should be a tool to retrieve that from comments (or actual signatures?), resulting in either HTML or Markdown documentation that can be automatically integrated in the "manual-style" documentation. Best Urs
Re: openLilyLib website
On 02/22, Urs Liska wrote: > Am Freitag, den 21.02.2020, 18:52 -0800 schrieb ma...@masonhock.com: > > Do you want something similar [Sphinx][1] or [Doxygen][2]? > > Yes, something similar, but IISC these tools (which I had of course > taken notice of earlier) are limited to their set of languages. Yes, that's right. I wasn't recommending those tools, just confirming that you are looking for something similar. > Craig Dabelstein did some research and suggested NaturalDocs ( > https://www.naturaldocs.org/) which looks nice and suitably > extensible/configurable for our purpose. I'm a little bit wary about > the fact that it's a project with a single contributor, and I know what > I'm talking about ... > (also I *think* it would be good to have a Python-based tool because it > might be easier to extend and to integrate with the other stuff. > > > Are the > > source files in Scheme, Lilypond, or a combination of both? > > Of course both. > The end-user interface is usually LilyPond, but the functionality is > all LilyPond, Scheme-in-LilyPond, and Scheme-in-Scheme-files. Huh. I thought it would be easier to find a good documentation generator for Scheme, but after searching for a while everything I found appears to have gone unmaintained for at least a few years (which I'm guessing is what you're afraid could happen with NaturalDocs?), and other than NaturalDocs I couldn't find anything that supports configuration for arbitrary languages. signature.asc Description: PGP signature
Re: openLilyLib website
Urs Liska writes: > Hi all, > > as a starting point for a - hopefully - comprehensive documentation > effort I have finally updated https://openlilylib.org with a completely > new website, which I'd like to have some feedback about and > contributions for. > > There are several parts to that effort, most of which are essentially > not started yet. > > * A general introduction website. This is basically complete and >should finally give a proper introduction about what OLL "is" and >how it can be made to work > * Independent sub-sites for each OLL package. These have not been >written at all, only the links to empty starting pages work without >404 errors. > * I've settled with MkDocs (https://www.mkdocs.org), which seems to >provide what I need, especially a suitable way to hook into and >extend to our needs. > * Each sub-site is maintained in a separate Git repository and >included as a Git submodule, so it should be straightforward to >manage independent authoring of the documentation by the respective >package maintainers. > * There's a link to a contributor's guide, which is also essentially >empty, except for an entry page. > > What I have so far is an infrastructure for textual, Markdown-authored > manuals, although I have already created a plugin for LilyPond syntax > highlighting using python-ly ( > https://github.com/uliska/markdown-lilypond/). > > What I really *want* to have but have no idea so far how to achieve is > additional code/API documentation retrieved from the actual source > files. There should be a tool to retrieve that from comments (or actual > signatures?), resulting in either HTML or Markdown documentation that > can be automatically integrated in the "manual-style" documentation. With regard to the LilyPond-book/Texinfo route, it might be worth considering that Asciidoc (which Git documentation is written in) can be converted via the route (grabbed lines from the git Documentation Makefile): user-manual.texi: user-manual.xml $(QUIET_DB2TEXI)$(RM) $@+ $@ && \ $(DOCBOOK2X_TEXI) user-manual.xml --encoding=UTF-8 --to-stdout >$@++ && \ $(PERL_PATH) fix-texi.perl <$@++ >$@+ && \ rm $@++ && \ mv $@+ $@ user-manual.xml: user-manual.txt user-manual.conf $(QUIET_ASCIIDOC)$(RM) $@+ $@ && \ $(TXT_TO_XML) -d book -o $@+ $< && \ mv $@+ $@ ASCIIDOC = asciidoc ASCIIDOC_EXTRA = ASCIIDOC_HTML = xhtml11 ASCIIDOC_DOCBOOK = docbook ASCIIDOC_CONF = -f asciidoc.conf ASCIIDOC_COMMON = $(ASCIIDOC) $(ASCIIDOC_EXTRA) $(ASCIIDOC_CONF) \ -agit_version=$(GIT_VERSION) TXT_TO_HTML = $(ASCIIDOC_COMMON) -b $(ASCIIDOC_HTML) TXT_TO_XML = $(ASCIIDOC_COMMON) -b $(ASCIIDOC_DOCBOOK) DOCBOOK2X_TEXI = docbook2x-texi So that's a bunch of stuff that can convert via Docbook XML to Texinfo. And it's likely that some Wiki-acceptable format convertable to Docbook XML exists. -- David Kastrup My replies have a tendency to cause friction. To help mitigating damage, feel free to forward problematic posts to me adding a subject like "timeout 1d" (for a suggested timeout of 1 day) or "offensive".
Re: openLilyLib website
Hi Simon, Am Sonntag, den 23.02.2020, 18:42 +0100 schrieb Simon Albrecht: > Hi Urs, > > this looks great! As often, I can’t offer much more than applause > and > some nitpicking… Well, that's useful too. > > On the ‘About LilyPond’ I’d rather write “GNU LilyPond is a music > notation program /strongly inspired by/ traditional craftsmanship”. > After all, the process of working with LilyPond doesn’t resemble > traditional craftsmanship at all… Good point. > > The use of the word ‘domain’ in the About/LilyPond and /openLilyLib > seems quite technical considering IIUC the site is supposed to be > read > also by people without a background in software development… I must say I'm pretty unhappy with that whole page. I actually replaced the bullet list with some text just to be able to send that email to the list. The introduction on that page should be even shorter, but at the same time more friendly, casual, positively inspiring than the current text tending to be too technical and boring. > > ‘Get started/Install openLilyLib’ has an instance of [oll-core} that > seems unintentional. Thanks, had already been spotted and fixed (although not uploaded yet). > > s/documenation/documentation Thanks. Best Urs > > … but generally, the texts seem very appropriate and useful :-) > > Very best regards > Simon > > On 20.02.20 22:44, Urs Liska wrote: > > Hi all, > > > > as a starting point for a - hopefully - comprehensive documentation > > effort I have finally updated https://openlilylib.org with a > > completely > > new website, which I'd like to have some feedback about and > > contributions for. > > > > There are several parts to that effort, most of which are > > essentially > > not started yet. > > > > * A general introduction website. This is basically complete and > > should finally give a proper introduction about what OLL "is" > > and > > how it can be made to work > > * Independent sub-sites for each OLL package. These have not been > > written at all, only the links to empty starting pages work > > without > > 404 errors. > > * I've settled with MkDocs (https://www.mkdocs.org), which seems > > to > > provide what I need, especially a suitable way to hook into and > > extend to our needs. > > * Each sub-site is maintained in a separate Git repository and > > included as a Git submodule, so it should be straightforward to > > manage independent authoring of the documentation by the > > respective > > package maintainers. > > * There's a link to a contributor's guide, which is also > > essentially > > empty, except for an entry page. > > > > What I have so far is an infrastructure for textual, Markdown- > > authored > > manuals, although I have already created a plugin for LilyPond > > syntax > > highlighting using python-ly ( > > https://github.com/uliska/markdown-lilypond/). > > > > What I really *want* to have but have no idea so far how to achieve > > is > > additional code/API documentation retrieved from the actual source > > files. There should be a tool to retrieve that from comments (or > > actual > > signatures?), resulting in either HTML or Markdown documentation > > that > > can be automatically integrated in the "manual-style" > > documentation. > > > > Best > > Urs > > > >
Re: openLilyLib website
Hi David, Am Sonntag, den 23.02.2020, 22:12 +0100 schrieb David Kastrup: > Urs Liska writes: > > > Hi all, > > > > as a starting point for a - hopefully - comprehensive documentation > > effort I have finally updated https://openlilylib.org with a > > completely > > new website, which I'd like to have some feedback about and > > contributions for. > > > > There are several parts to that effort, most of which are > > essentially > > not started yet. > > > > * A general introduction website. This is basically complete and > >should finally give a proper introduction about what OLL "is" > > and > >how it can be made to work > > * Independent sub-sites for each OLL package. These have not been > >written at all, only the links to empty starting pages work > > without > >404 errors. > > * I've settled with MkDocs (https://www.mkdocs.org), which seems > > to > >provide what I need, especially a suitable way to hook into and > >extend to our needs. > > * Each sub-site is maintained in a separate Git repository and > >included as a Git submodule, so it should be straightforward to > >manage independent authoring of the documentation by the > > respective > >package maintainers. > > * There's a link to a contributor's guide, which is also > > essentially > >empty, except for an entry page. > > > > What I have so far is an infrastructure for textual, Markdown- > > authored > > manuals, although I have already created a plugin for LilyPond > > syntax > > highlighting using python-ly ( > > https://github.com/uliska/markdown-lilypond/). > > > > What I really *want* to have but have no idea so far how to achieve > > is > > additional code/API documentation retrieved from the actual source > > files. There should be a tool to retrieve that from comments (or > > actual > > signatures?), resulting in either HTML or Markdown documentation > > that > > can be automatically integrated in the "manual-style" > > documentation. > I must admit I don't fully understand what your comments are actually targeted at. > With regard to the LilyPond-book/Texinfo route, it might be worth > considering that Asciidoc (which Git documentation is written in) can > be > converted via the route (grabbed lines from the git Documentation > Makefile): > > user-manual.texi: user-manual.xml > $(QUIET_DB2TEXI)$(RM) $@+ $@ && \ > $(DOCBOOK2X_TEXI) user-manual.xml --encoding=UTF-8 --to-stdout > >$@++ && \ > $(PERL_PATH) fix-texi.perl <$@++ >$@+ && \ > rm $@++ && \ > mv $@+ $@ > > user-manual.xml: user-manual.txt user-manual.conf > $(QUIET_ASCIIDOC)$(RM) $@+ $@ && \ > $(TXT_TO_XML) -d book -o $@+ $< && \ > mv $@+ $@ > > ASCIIDOC = asciidoc > ASCIIDOC_EXTRA = > ASCIIDOC_HTML = xhtml11 > ASCIIDOC_DOCBOOK = docbook > ASCIIDOC_CONF = -f asciidoc.conf > ASCIIDOC_COMMON = $(ASCIIDOC) $(ASCIIDOC_EXTRA) $(ASCIIDOC_CONF) \ > -agit_version=$(GIT_VERSION) > TXT_TO_HTML = $(ASCIIDOC_COMMON) -b $(ASCIIDOC_HTML) > TXT_TO_XML = $(ASCIIDOC_COMMON) -b $(ASCIIDOC_DOCBOOK) > > DOCBOOK2X_TEXI = docbook2x-texi > > So that's a bunch of stuff that can convert via Docbook XML to > Texinfo. > And it's likely that some Wiki-acceptable format convertable to > Docbook > XML exists. I've had a look at the Asciidoc homepage, and it looks like a tool I might like. But at this point we are talking about HTML targets (possibly PDF too), Markdown files and documentation generated from LilyPond/Scheme files. Nothing about DocBook, texidoc etc. What are you actually intending to convey with these Makefile excerpts? That I should consider AsciiDoc as a possible platform for the documentation? What would be in there that makes it a (better and) suitable tool for the task? (This is not meant as an objection [somewhat hard to express], rather a request for clarification. I need more a more concrete idea what you want me to consider.) Urs >
Re: openLilyLib website
Urs Liska writes: > Hi David, > > Am Sonntag, den 23.02.2020, 22:12 +0100 schrieb David Kastrup: >> Urs Liska writes: >> >> > Hi all, >> > >> > as a starting point for a - hopefully - comprehensive documentation >> > effort I have finally updated https://openlilylib.org with a >> > completely >> > new website, which I'd like to have some feedback about and >> > contributions for. >> > >> > There are several parts to that effort, most of which are >> > essentially >> > not started yet. >> > >> > * A general introduction website. This is basically complete and >> >should finally give a proper introduction about what OLL "is" >> > and >> >how it can be made to work >> > * Independent sub-sites for each OLL package. These have not been >> >written at all, only the links to empty starting pages work >> > without >> >404 errors. >> > * I've settled with MkDocs (https://www.mkdocs.org), which seems >> > to >> >provide what I need, especially a suitable way to hook into and >> >extend to our needs. >> > * Each sub-site is maintained in a separate Git repository and >> >included as a Git submodule, so it should be straightforward to >> >manage independent authoring of the documentation by the >> > respective >> >package maintainers. >> > * There's a link to a contributor's guide, which is also >> > essentially >> >empty, except for an entry page. >> > >> > What I have so far is an infrastructure for textual, Markdown- >> > authored >> > manuals, although I have already created a plugin for LilyPond >> > syntax >> > highlighting using python-ly ( >> > https://github.com/uliska/markdown-lilypond/). >> > >> > What I really *want* to have but have no idea so far how to achieve >> > is >> > additional code/API documentation retrieved from the actual source >> > files. There should be a tool to retrieve that from comments (or >> > actual >> > signatures?), resulting in either HTML or Markdown documentation >> > that >> > can be automatically integrated in the "manual-style" >> > documentation. >> > > I must admit I don't fully understand what your comments are actually > targeted at. > >> With regard to the LilyPond-book/Texinfo route, it might be worth >> considering that Asciidoc (which Git documentation is written in) can >> be >> converted via the route (grabbed lines from the git Documentation >> Makefile): >> >> user-manual.texi: user-manual.xml >> $(QUIET_DB2TEXI)$(RM) $@+ $@ && \ >> $(DOCBOOK2X_TEXI) user-manual.xml --encoding=UTF-8 --to-stdout >> >$@++ && \ >> $(PERL_PATH) fix-texi.perl <$@++ >$@+ && \ >> rm $@++ && \ >> mv $@+ $@ >> >> user-manual.xml: user-manual.txt user-manual.conf >> $(QUIET_ASCIIDOC)$(RM) $@+ $@ && \ >> $(TXT_TO_XML) -d book -o $@+ $< && \ >> mv $@+ $@ >> >> ASCIIDOC = asciidoc >> ASCIIDOC_EXTRA = >> ASCIIDOC_HTML = xhtml11 >> ASCIIDOC_DOCBOOK = docbook >> ASCIIDOC_CONF = -f asciidoc.conf >> ASCIIDOC_COMMON = $(ASCIIDOC) $(ASCIIDOC_EXTRA) $(ASCIIDOC_CONF) \ >> -agit_version=$(GIT_VERSION) >> TXT_TO_HTML = $(ASCIIDOC_COMMON) -b $(ASCIIDOC_HTML) >> TXT_TO_XML = $(ASCIIDOC_COMMON) -b $(ASCIIDOC_DOCBOOK) >> >> DOCBOOK2X_TEXI = docbook2x-texi >> >> So that's a bunch of stuff that can convert via Docbook XML to >> Texinfo. >> And it's likely that some Wiki-acceptable format convertable to >> Docbook >> XML exists. > > I've had a look at the Asciidoc homepage, and it looks like a tool I > might like. > > But at this point we are talking about HTML targets (possibly PDF too), > Markdown files and documentation generated from LilyPond/Scheme files. > Nothing about DocBook, texidoc etc. > > What are you actually intending to convey with these Makefile excerpts? > That I should consider AsciiDoc as a possible platform for the > documentation? What would be in there that makes it a (better and) > suitable tool for the task? > (This is not meant as an objection [somewhat hard to express], rather a > request for clarification. I need more a more concrete idea what you > want me to consider.) Basically this boils down to pointing out that Texinfo, which has some prominence in the LilyPond core, has conversion paths via docbook2x-texi from formats that are more akin to what one would use automated documentation extraction on. That's all. I have not enough clue about openLilyLib to figure out whether that information might be useful for you. I just threw that out and it would seem like I made it sound like more (or something different) than it actually is. Just some keyword-triggered info that may or may not be interesting for whatever this conversation I barreled in was about. Sorry if it's just nonsense and occupied too much of your time trying to make sense of. David -- David Kastrup My replies have a tendency to cause friction. To help mitigating damage, feel free to forward problematic posts to me adding a subject like "timeout 1d" (for a suggested timeout of 1 day) or "offensive"