Loading ...
Sorry, an error occurred while loading the content.

Two New Software-Related Ideas

Expand Messages
  • Shlomi Fish
    Hi all! You can find two new software-related ideas that I scribbled down here: * http://www.shlomifish.org/philosophy/ideas/#tucan *
    Message 1 of 12 , Sep 19, 2006
    View Source
    • 0 Attachment
      Hi all!

      You can find two new software-related ideas that I scribbled down here:

      * http://www.shlomifish.org/philosophy/ideas/#tucan

      * http://www.shlomifish.org/philosophy/ideas/unixdoc/

      Please let me know what you think.

      Regards,

      Shlomi Fish

      ---------------------------------------------------------------------
      Shlomi Fish shlomif@...
      Homepage: http://www.shlomifish.org/

      Chuck Norris wrote a complete Perl 6 implementation in a day but then
      destroyed all evidence with his bare hands, so no one will know his secrets.
    • Nadav Har'El
      ... Once upon a time (about a decade ago), I used tkman to read manual pages. It was similar to what you were looking for - a graphical, hypertext manual
      Message 2 of 12 , Sep 19, 2006
      View Source
      • 0 Attachment
        On Tue, Sep 19, 2006, Shlomi Fish wrote about "[hackers-il] Two New Software-Related Ideas":
        > You can find two new software-related ideas that I scribbled down here:
        >
        > * http://www.shlomifish.org/philosophy/ideas/unixdoc/

        Once upon a time (about a decade ago), I used "tkman" to read manual pages.
        It was similar to what you were looking for - a graphical, hypertext manual
        page (and now I see, also texinfo) browser. It added links between pages
        (so you could click on those "see also" sections or mentions of cat(1)),
        and was much more "cool" than using "man" in a text terminal.

        --
        Nadav Har'El | Tuesday, Sep 19 2006, 26 Elul 5766
        nyh@... |-----------------------------------------
        Phone +972-523-790466, ICQ 13349191 |Why do programmers mix up Christmas and
        http://nadav.harel.org.il |Halloween? Because DEC 25 = OCT 31
      • Chen Shapira
        Regarding Tucan: Do you have any explicit requirements? Because I don t see what exactly is the problem with Parrot and how YAML and dbus can solve the
        Message 3 of 12 , Sep 21, 2006
        View Source
        • 0 Attachment
          Regarding Tucan:
          Do you have any explicit requirements? Because I don't see what exactly
          is the problem with Parrot and how YAML and dbus can solve the problem.

          I didn't work with any of the APIs you mentioned, but I have experience
          with message buses. Buses are a cool way to communicate between
          different programs regardless of the language, but they can't enable you
          to use code in other languages any more that HTTP can. It is just a high
          level communication protocol that can be very useful in some specific
          cases.

          Moreover, I believe that asking the source and target languages to be
          able to convert to and from a common language is not something you can
          work around. It will be at the very core of any solution you find, if
          you don't want Parrot's byte code, you'll find yourself compiling code
          into XML or bus messages, and personally I prefer bytecode to XML any
          day.


          > -----Original Message-----
          > From: hackers-il@yahoogroups.com [mailto:hackers-il@yahoogroups.com]
          On
          > Behalf Of Shlomi Fish
          > Sent: Tuesday, September 19, 2006 10:11 AM
          > To: Hackers-IL,
          > Subject: [hackers-il] Two New Software-Related Ideas
          >
          > Hi all!
          >
          > You can find two new software-related ideas that I scribbled down
          here:
          >
          > * http://www.shlomifish.org/philosophy/ideas/#tucan
          >
          > * http://www.shlomifish.org/philosophy/ideas/unixdoc/
          >
          > Please let me know what you think.
          >
          > Regards,
          >
          > Shlomi Fish
          >
          > ---------------------------------------------------------------------
          > Shlomi Fish shlomif@...
          > Homepage: http://www.shlomifish.org/
          >
          > Chuck Norris wrote a complete Perl 6 implementation in a day but then
          > destroyed all evidence with his bare hands, so no one will know his
          > secrets.
          >
          >
          >
          > Yahoo! Groups Links
          >
          >
          >
          >
          >
          >
          >


          ______________________________________________________________________
          This email has been scanned by the MessageLabs Email Security System.
          For more information please visit http://www.messagelabs.com/email
          ______________________________________________________________________
        • Beni Cherniavsky
          ... I think the common format fed to presentation engines should be HTML (perhaps adding some microformats to add semantics). It satisfies all your goals,
          Message 4 of 12 , Sep 21, 2006
          View Source
          • 0 Attachment
            On 19/09/06, Nadav Har'El <nyh@...> wrote:
            > On Tue, Sep 19, 2006, Shlomi Fish wrote about "[hackers-il] Two New Software-Related Ideas":
            > > You can find two new software-related ideas that I scribbled down here:
            > >
            > > * http://www.shlomifish.org/philosophy/ideas/unixdoc/
            >
            > Once upon a time (about a decade ago), I used "tkman" to read manual pages.
            > It was similar to what you were looking for - a graphical, hypertext manual
            > page (and now I see, also texinfo) browser. It added links between pages
            > (so you could click on those "see also" sections or mentions of cat(1)),
            > and was much more "cool" than using "man" in a text terminal.
            >
            I think the common format fed to presentation engines should be HTML
            (perhaps adding some microformats to add semantics). It satisfies all
            your goals, with more tools around it than any other format. links
            does a decent job of showing HTML at the terminal.

            Existing {texi,info,man}2html converters are a bit lame from the
            output aesthetics perspective. Nevertheless, 'dwww' (debian only I
            think) does a very nice job of integrating them into a uniform
            browsing experience.

            --
            Beni Cherniavsky <cben@...>, who can only read email on weekends.
          • Shlomi Fish
            ... I see. Nevertheless, I need such a thing also for other formats, and with other features. And naturally, there should be a terminal front-end. Regards,
            Message 5 of 12 , Sep 22, 2006
            View Source
            • 0 Attachment
              On Tuesday 19 September 2006 13:16, Nadav Har'El wrote:
              > On Tue, Sep 19, 2006, Shlomi Fish wrote about "[hackers-il] Two New
              Software-Related Ideas":
              > > You can find two new software-related ideas that I scribbled down here:
              > >
              > > * http://www.shlomifish.org/philosophy/ideas/unixdoc/
              >
              > Once upon a time (about a decade ago), I used "tkman" to read manual pages.
              > It was similar to what you were looking for - a graphical, hypertext manual
              > page (and now I see, also texinfo) browser. It added links between pages
              > (so you could click on those "see also" sections or mentions of cat(1)),
              > and was much more "cool" than using "man" in a text terminal.

              I see. Nevertheless, I need such a thing also for other formats, and with
              other features. And naturally, there should be a terminal front-end.

              Regards,

              Shlomi Fish

              ---------------------------------------------------------------------
              Shlomi Fish shlomif@...
              Homepage: http://www.shlomifish.org/

              Chuck Norris wrote a complete Perl 6 implementation in a day but then
              destroyed all evidence with his bare hands, so no one will know his secrets.
            • Shlomi Fish
              ... Parrot is a high-level virtual machine, the aim of which is that one can implement dynamic languages (Perl, Python, Tcl, Ruby, PHP, etc.) on top of with
              Message 6 of 12 , Sep 22, 2006
              View Source
              • 0 Attachment
                On Thursday 21 September 2006 13:43, Chen Shapira wrote:
                > Regarding Tucan:
                > Do you have any explicit requirements? Because I don't see what exactly
                > is the problem with Parrot and how YAML and dbus can solve the problem.

                Parrot is a high-level virtual machine, the aim of which is that one can
                implement dynamic languages (Perl, Python, Tcl, Ruby, PHP, etc.) on top of
                with ease. Assuming all languages are implemented above Parrot, then the
                intention of the Parrot hackers it that it would be easy to call from one
                language to another.

                The downside of Parrot is that it does little to bridge perl5, CPython, CRuby,
                etc. All of these language do not yet work above Python and after they do,
                they will lose a lot of their existing C-based APIs.

                My (pie-in-the-sky) intention with Tucan is that I can do the following in
                Python: (not tested but should be OK)

                <<<<<<<<<<<<<<
                import tucan.perl5.HTML.Widgets.NavMenu

                nav_menu = HTML.Widgets.NavMenu.new(
                'path_info': "/me/",
                'current_host': "default",
                'hosts' : { 'default':{"base_url":"http://www.hello.com/"}},
                'tree_contents': ...
                )

                results = nav_menu.render()

                nav_menu_html = "\n".join(results['html'])
                >>>>>>>>>>>>>>

                And Tucan will do the right thing and interface between Python and the Perl
                based http://search.cpan.org/dist/HTML-Widgets-NavMenu/.

                >
                > I didn't work with any of the APIs you mentioned, but I have experience
                > with message buses. Buses are a cool way to communicate between
                > different programs regardless of the language, but they can't enable you
                > to use code in other languages any more that HTTP can. It is just a high
                > level communication protocol that can be very useful in some specific
                > cases.

                Yes, I realise that. YAML is a common format for storing nested data
                structures. dbus is a lightweight communication protocol.

                >
                > Moreover, I believe that asking the source and target languages to be
                > able to convert to and from a common language is not something you can
                > work around. It will be at the very core of any solution you find, if
                > you don't want Parrot's byte code, you'll find yourself compiling code
                > into XML or bus messages, and personally I prefer bytecode to XML any
                > day.

                I'm not talking about compiling code into anything. The Perl 5 VM will handle
                the Perl 5 code, and the Python VM will handle the VM one. Tucan will load
                all the VMs on demand. What it will do is transfer calls that Perl 5 code to
                methods written in Python to the Python VM, after which their results will be
                returned to the Perl 5 code.

                This way every language can make use of code written in any other language.

                Regards,

                Shlomi Fish

                >
                > > -----Original Message-----
                > > From: hackers-il@yahoogroups.com [mailto:hackers-il@yahoogroups.com]
                >
                > On
                >
                > > Behalf Of Shlomi Fish
                > > Sent: Tuesday, September 19, 2006 10:11 AM
                > > To: Hackers-IL,
                > > Subject: [hackers-il] Two New Software-Related Ideas
                > >
                > > Hi all!
                > >
                > > You can find two new software-related ideas that I scribbled down
                >
                > here:
                > > * http://www.shlomifish.org/philosophy/ideas/#tucan
                > >
                > > * http://www.shlomifish.org/philosophy/ideas/unixdoc/
                > >
                > > Please let me know what you think.
                > >
                > > Regards,
                > >
                > > Shlomi Fish
                > >
                > > ---------------------------------------------------------------------
                > > Shlomi Fish shlomif@...
                > > Homepage: http://www.shlomifish.org/
                > >
                > > Chuck Norris wrote a complete Perl 6 implementation in a day but then
                > > destroyed all evidence with his bare hands, so no one will know his
                > > secrets.
                > >
                > >
                > >
                > > Yahoo! Groups Links
                >
                > ______________________________________________________________________
                > This email has been scanned by the MessageLabs Email Security System.
                > For more information please visit http://www.messagelabs.com/email
                > ______________________________________________________________________

                --

                ---------------------------------------------------------------------
                Shlomi Fish shlomif@...
                Homepage: http://www.shlomifish.org/

                Chuck Norris wrote a complete Perl 6 implementation in a day but then
                destroyed all evidence with his bare hands, so no one will know his secrets.
              • Omer Zak
                One thing whose absence I sorely feel: An improved way to have apropos. I often need to accomplish a task, remember that there is some command which can be
                Message 7 of 12 , Oct 2, 2006
                View Source
                • 0 Attachment
                  One thing whose absence I sorely feel:
                  An improved way to have apropos.

                  I often need to accomplish a task, remember that there is some command
                  which can be used to accomplish this task. But do not remember which
                  command - not because I was lazy, did not RTFM, or whatever - but
                  because my need to perform this specific task is rare enough for me not
                  to remember how to accomplish it.

                  It would be nice to have some sort of database or a search engine, into
                  which you can type keywords describing the task and it displays to you a
                  list of potentially relevant software tools, command line options and
                  commands which might be relevant to performance of that task.

                  --- Omer


                  On Tue, 2006-09-19 at 13:16 +0300, Nadav Har'El wrote:
                  > On Tue, Sep 19, 2006, Shlomi Fish wrote about "[hackers-il] Two New Software-Related Ideas":
                  > > You can find two new software-related ideas that I scribbled down here:
                  > >
                  > > * http://www.shlomifish.org/philosophy/ideas/unixdoc/
                  >
                  > Once upon a time (about a decade ago), I used "tkman" to read manual pages.
                  > It was similar to what you were looking for - a graphical, hypertext manual
                  > page (and now I see, also texinfo) browser. It added links between pages
                  > (so you could click on those "see also" sections or mentions of cat(1)),
                  > and was much more "cool" than using "man" in a text terminal.
                  --
                  You haven't made an impact on the world before you caused a Debian
                  release to be named after Snufkin.
                  My own blog is at http://tddpirate.livejournal.com/

                  My opinions, as expressed in this E-mail message, are mine alone.
                  They do not represent the official policy of any organization with which
                  I may be affiliated in any way.
                  WARNING TO SPAMMERS: at http://www.zak.co.il/spamwarning.html
                • Beni Cherniavsky
                  ... [...] ... On debian, I frequently use apt-cache search . When I don t know how to formulate the search I launch synaptic and try all the searches I can
                  Message 8 of 12 , Oct 2, 2006
                  View Source
                  • 0 Attachment
                    On 02/10/06, Omer Zak <omerz@...> wrote:
                    > One thing whose absence I sorely feel:
                    > An improved way to have apropos.
                    [...]
                    > It would be nice to have some sort of database or a search engine, into
                    > which you can type keywords describing the task and it displays to you a
                    > list of potentially relevant software tools, command line options and
                    > commands which might be relevant to performance of that task.
                    >
                    On debian, I frequently use "apt-cache search". When I don't know how to
                    formulate the search I launch "synaptic" and try all the searches I
                    can think of.
                    Both search the full description texts of all packages, which gives better
                    results than man -k. Also, if I'm looking for the right tool, I want
                    to consider
                    all availiable packages, not only those those that happen to be installed.

                    On other distributions, I expect that searching the package database will
                    also give good results.

                    --
                    Beni Cherniavsky <cben@...>
                  • Amit Aronovitch
                    This is a repost of my message from 21/9 (I was told it was rejected and I have to repost) comments relate to the unixdoc idea. rants--well, guess I was in a
                    Message 9 of 12 , Oct 24, 2006
                    View Source
                    • 0 Attachment
                      This is a repost of my message from 21/9
                      (I was told it was rejected and I have to repost)

                      comments relate to the "unixdoc" idea.

                      rants--well, guess I was in a cynical mood that day.

                      ------------------------------------------------------------------------------------------
                      From: Amit Aronovitch <aronovitch@...> Mailed-By: gmail.com
                      To: hackers-il@yahoogroups.com
                      Date: Sep 21, 2006 1:42 AM
                      Subject: Re: [hackers-il] Two New Software-Related Ideas

                      (a)

                      http://www.gnu.org/software/texinfo/

                      You all probably know that, but shame it is not mentioned - not even
                      as an input format.

                      Few facts (I know a few people which I was very surprised to find out
                      they did not even know that info exists):

                      1) All the detailed/updated documentation of all GNU components come
                      as info. man pages are either a brief summary or inserted by the
                      distros) - try "info bash", "info find", "info coreutils" on any GNU
                      based system (in fact all propriatary unices I happened to work on in
                      the last 5 years offer a provider-prepared compilation of standard gnu
                      tools - info included)

                      (Better to avoid using man on GNU tools - in many cases important
                      information was dropped out. Plus - once you get used to the interface
                      - navigation and searching is much more flexible - even for reading
                      plain manpages)

                      2) The texinfo system is menu driven and hyperlinked. It supports
                      generating a pretty hardcopy (using tex) and html.

                      3) The major problem is that the standard browser (info info -
                      contains a tutorial), is very emacs-ish, which, while immediate and
                      intuitive to some of us, is intimidating to others. Also it's not new
                      - therefore not "cool" (people seem to think that anything that was
                      invented earlier than last year is old-fashioned and can't be any
                      good).
                      However: nothing prevents you for writing graphical viewers. In fact,

                      4) Modern desktop systems already provide interface to the info system
                      (in Gnome yelp, you find the info directory under "command line help"
                      next to the man pages).

                      ---------------------------

                      (b)

                      Debian provides a nice set of tools, which should let you read all
                      docs (man, info and html-docs) of installed packages using your web
                      browser ( I think most packages do register themself using doc-base) .
                      I believe the 3 alternatives are "dwww","doc-central" and "dhelp"
                      (each having different set of capabilities and different requirements
                      - e.g. dhelp does not require a webserver)
                      This is all I can say right now - since I did not yet try any of the 3.

                      ----------------------------

                      (c)

                      Some rants about "modern style" of documentation (Gnome/KDE docbooks) :
                      The modern desktops seem to think that their form of docs is so much
                      better than the "old fashioned" man/info, that they don't bother
                      providing any other form, even for commandline tools. It seems that
                      even Debian packagers (which are usually very strict about having
                      manpages) fall to that dogma.

                      However, the documentation that does exist in their fancy graphical
                      system (which of course very *cool* and *modern*), is close to 100%
                      what I call "useless menu listing"(*) - all the trivial parts that you
                      can easily find out yourself are documented, but anything useful
                      enough you'd bother opening the help for is considered "expert-level
                      details" that would only confuse and bother the poor stupid^H Eh...
                      sorry... *respected* user in Joelspeak (
                      http://www.joelonsoftware.com/uibook/chapters/fog0000000062.html ).
                      I say - if you really think "users" (FTR, I resent the term) don't
                      read the manual (obviously false generalization), don't document at
                      all (see if you can get you clients to accept "users don't read" as an
                      excuse for that...). If you bother to document - document the
                      *important* stuff. If I clicked "help" I probably *do* want to get
                      some info - right?

                      AA

                      (*) Go to the nearest book-shop and pick some random computer book,
                      say a (fictitional) 600 pages "F. Mbogo's Absolute Guide to Microsoft
                      Word", open some random page.
                      Most probably you'll see something like this:
                      ...
                      1.2.3 The File menu:
                      1.2.3.1 The "New" command
                      Use this to start working on a new document. A menu pops up asking
                      you to choose the type of document - see figure 1.16.
                      1.2.3.2 The "Open" command
                      To open a new file select this option. A window will pop up,
                      allowing you to pick the file you want to edit (see figure 1.17).
                      Detailed description on using that window can be found in F. Mbogo's
                      "Power User's Guide to Windows Explorer"
                      ...

                      On 9/19/06, Shlomi Fish <shlomif@...> wrote:
                      > Hi all!
                      >
                      > You can find two new software-related ideas that I scribbled down here:
                      >
                      > * http://www.shlomifish.org/philosophy/ideas/#tucan
                      >
                      > * http://www.shlomifish.org/philosophy/ideas/unixdoc/
                      >
                      > Please let me know what you think.
                      >
                      > Regards,
                      >
                      > Shlomi Fish
                      >
                    • Shlomi Fish
                      ... Yes, I m aware of GNU texinfo. I don t remember why I didn t include it. Perhaps it was out of due frustration from its limitations or the fact that I
                      Message 10 of 12 , Oct 25, 2006
                      View Source
                      • 0 Attachment
                        On Wednesday 25 October 2006 01:19, Amit Aronovitch wrote:
                        > This is a repost of my message from 21/9
                        > (I was told it was rejected and I have to repost)
                        >
                        > comments relate to the "unixdoc" idea.
                        >
                        > rants--well, guess I was in a cynical mood that day.
                        >
                        > ---------------------------------------------------------------------------
                        >--------------- From: Amit Aronovitch <aronovitch@...> Mailed-By:
                        > gmail.com
                        > To: hackers-il@yahoogroups.com
                        > Date: Sep 21, 2006 1:42 AM
                        > Subject: Re: [hackers-il] Two New Software-Related Ideas
                        >
                        > (a)
                        >
                        > http://www.gnu.org/software/texinfo/
                        >
                        > You all probably know that, but shame it is not mentioned - not even
                        > as an input format.

                        Yes, I'm aware of GNU texinfo. I don't remember why I didn't include it.
                        Perhaps it was out of due frustration from its limitations or the fact that I
                        found the "info" reader to be so counter-intuitive.

                        I normally use the HTML documentation (as converted from the texinfo sources)
                        at gnu.org.

                        >
                        > Few facts (I know a few people which I was very surprised to find out
                        > they did not even know that info exists):
                        >
                        > 1) All the detailed/updated documentation of all GNU components come
                        > as info. man pages are either a brief summary or inserted by the
                        > distros) - try "info bash", "info find", "info coreutils" on any GNU
                        > based system (in fact all propriatary unices I happened to work on in
                        > the last 5 years offer a provider-prepared compilation of standard gnu
                        > tools - info included)
                        >
                        > (Better to avoid using man on GNU tools - in many cases important
                        > information was dropped out. Plus - once you get used to the interface
                        > - navigation and searching is much more flexible - even for reading
                        > plain manpages)

                        One problem Nadav once mentioned with the info pages, is that they are split
                        into several sections and sub-sections and so one cannot search them for a
                        keyword as conveniently as one can using a man page.

                        >
                        > 2) The texinfo system is menu driven and hyperlinked. It supports
                        > generating a pretty hardcopy (using tex) and html.

                        Yes, albeit its support for hypertext is relatively limited in comparison to
                        DocBook.

                        >
                        > 3) The major problem is that the standard browser (info info -
                        > contains a tutorial), is very emacs-ish, which, while immediate and
                        > intuitive to some of us, is intimidating to others.

                        I could never get the hang of the "info" command line. :-)

                        > Also it's not new
                        > - therefore not "cool" (people seem to think that anything that was
                        > invented earlier than last year is old-fashioned and can't be any
                        > good).
                        > However: nothing prevents you for writing graphical viewers. In fact,
                        >
                        > 4) Modern desktop systems already provide interface to the info system
                        > (in Gnome yelp, you find the info directory under "command line help"
                        > next to the man pages).
                        >

                        Yes.

                        > ---------------------------
                        >
                        > (b)
                        >
                        > Debian provides a nice set of tools, which should let you read all
                        > docs (man, info and html-docs) of installed packages using your web
                        > browser ( I think most packages do register themself using doc-base) .
                        > I believe the 3 alternatives are "dwww","doc-central" and "dhelp"
                        > (each having different set of capabilities and different requirements
                        > - e.g. dhelp does not require a webserver)
                        > This is all I can say right now - since I did not yet try any of the 3.

                        Interesting. I suppose it uses texinfo2html and man2html. Not precisely the
                        level of integration that I'm looking for.

                        >
                        > ----------------------------
                        >
                        > (c)
                        >
                        > Some rants about "modern style" of documentation (Gnome/KDE docbooks) :
                        > The modern desktops seem to think that their form of docs is so much
                        > better than the "old fashioned" man/info, that they don't bother
                        > providing any other form, even for commandline tools. It seems that
                        > even Debian packagers (which are usually very strict about having
                        > manpages) fall to that dogma.
                        >
                        > However, the documentation that does exist in their fancy graphical
                        > system (which of course very *cool* and *modern*), is close to 100%
                        > what I call "useless menu listing"(*) - all the trivial parts that you
                        > can easily find out yourself are documented, but anything useful
                        > enough you'd bother opening the help for is considered "expert-level
                        > details" that would only confuse and bother the poor stupid^H Eh...
                        > sorry... *respected* user in Joelspeak (
                        > http://www.joelonsoftware.com/uibook/chapters/fog0000000062.html ).
                        > I say - if you really think "users" (FTR, I resent the term) don't
                        > read the manual (obviously false generalization), don't document at
                        > all (see if you can get you clients to accept "users don't read" as an
                        > excuse for that...). If you bother to document - document the
                        > *important* stuff. If I clicked "help" I probably *do* want to get
                        > some info - right?

                        You're right about that. Of course, I often end up consulting the code for how
                        to do something using a program.

                        Regards,

                        Shlomi Fish

                        >
                        > AA
                        >
                        > (*) Go to the nearest book-shop and pick some random computer book,
                        > say a (fictitional) 600 pages "F. Mbogo's Absolute Guide to Microsoft
                        > Word", open some random page.
                        > Most probably you'll see something like this:
                        > ...
                        > 1.2.3 The File menu:
                        > 1.2.3.1 The "New" command
                        > Use this to start working on a new document. A menu pops up asking
                        > you to choose the type of document - see figure 1.16.
                        > 1.2.3.2 The "Open" command
                        > To open a new file select this option. A window will pop up,
                        > allowing you to pick the file you want to edit (see figure 1.17).
                        > Detailed description on using that window can be found in F. Mbogo's
                        > "Power User's Guide to Windows Explorer"
                        > ...
                        >
                        > On 9/19/06, Shlomi Fish <shlomif@...> wrote:
                        > > Hi all!
                        > >
                        > > You can find two new software-related ideas that I scribbled down here:
                        > >
                        > > * http://www.shlomifish.org/philosophy/ideas/#tucan
                        > >
                        > > * http://www.shlomifish.org/philosophy/ideas/unixdoc/
                        > >
                        > > Please let me know what you think.
                        > >
                        > > Regards,
                        > >
                        > > Shlomi Fish

                        --

                        ---------------------------------------------------------------------
                        Shlomi Fish shlomif@...
                        Homepage: http://www.shlomifish.org/

                        Chuck Norris wrote a complete Perl 6 implementation in a day but then
                        destroyed all evidence with his bare hands, so no one will know his secrets.
                      • Amit Aronovitch
                        ... Driven away because of the frontend... exactly as I said in (3). info is only intuitive to Emacsers, and most of them read the info with the emacs frontend
                        Message 11 of 12 , Oct 26, 2006
                        View Source
                        • 0 Attachment
                          On 10/25/06, Shlomi Fish <shlomif@...> wrote:

                          > Yes, I'm aware of GNU texinfo. I don't remember why I didn't include it.
                          > Perhaps it was out of due frustration from its limitations or the fact that I
                          > found the "info" reader to be so counter-intuitive.
                          >

                          Driven away because of the frontend... exactly as I said in (3).
                          info is only intuitive to Emacsers, and most of them read the info
                          with the emacs frontend anyway.
                          I wonder if there's a "viper"-like keyboard scheme for info, or a
                          standalone vi-like reader. These might do wonders to the popularity of
                          the system.

                          Anyways, if you consider writing a new doc framework - supporting
                          info as input source would greatly increase the amount of docs
                          available to your system (the useful ones, at least ;-)).

                          > I normally use the HTML documentation (as converted from the texinfo sources)
                          > at gnu.org.
                          >

                          That's remote documentation - not available when your'e offline, and
                          may not match the versions you have installed on your machine (that
                          is, unless your distro takes care of installing these htmls locally in
                          a centralized way - e.g. dhelp and friends)

                          Maybe it would be better to use your desktop's help system - I think
                          both KDE and Gnome help are aware of the installed info docs and allow
                          you to browse and read them.

                          >
                          > One problem Nadav once mentioned with the info pages, is that they are split
                          > into several sections and sub-sections and so one cannot search them for a
                          > keyword as conveniently as one can using a man page.

                          I was told that too.
                          But this is FALSE (maybe it was true in old versions - I don't know).

                          Both normal search (s, prompts for input, like '/' in man) and the
                          electric-search (^s, search-as-you-type, like '/' in firefox) work
                          across sections (they view the whole topic as if it was one big
                          document).
                          btw, in the Emacs frontend, electric search is page-local, which is,
                          in fact, one of the reasons I prefer the standalone info reader.
                        • Beni Cherniavsky
                          ... ``info foo | less`` gives you the whole manual in one document. But this loses hyperlinks and gains nothing. As amit says, `Ctrl-S` and `/` work accross
                          Message 12 of 12 , Oct 27, 2006
                          View Source
                          • 0 Attachment
                            On 25/10/06, Shlomi Fish <shlomif@...> wrote:
                            > One problem Nadav once mentioned with the info pages, is that they are split
                            > into several sections and sub-sections and so one cannot search them for a
                            > keyword as conveniently as one can using a man page.
                            >
                            ``info foo | less`` gives you the whole manual in one document. But
                            this loses hyperlinks and gains nothing.

                            As amit says, `Ctrl-S` and `/` work accross sections in the standalone
                            `info` reader. `i` searches the index which is usually high-quality.

                            > >
                            > > 3) The major problem is that the standard browser (info info -
                            > > contains a tutorial), is very emacs-ish, which, while immediate and
                            > > intuitive to some of us, is intimidating to others.
                            >
                            > I could never get the hang of the "info" command line. :-)
                            >
                            `pinfo` is friendlier but less powerful. Browser access (e.g. dwww)
                            is probably your best option today.

                            > >
                            > > Debian provides a nice set of tools, which should let you read all
                            > > docs (man, info and html-docs) of installed packages using your web
                            > > browser ( I think most packages do register themself using doc-base) .
                            > > I believe the 3 alternatives are "dwww","doc-central" and "dhelp"
                            > > (each having different set of capabilities and different requirements
                            > > - e.g. dhelp does not require a webserver)
                            > > This is all I can say right now - since I did not yet try any of the 3.
                            >
                            > Interesting. I suppose it uses texinfo2html and man2html. Not precisely the
                            > level of integration that I'm looking for.
                            >
                            What's missing?

                            --
                            Beni Cherniavsky <cben@...> (I read email only on weekends)
                          Your message has been successfully submitted and would be delivered to recipients shortly.