Re: [Hyperbole-users] [bugs@gnu.support: Re: Need more usage examples (Was: Subscription within Hyperbole did not work, do I need to repeat sending bugs?)]

[Robert Weiner]

> On Oct 7, 2019, at 7:22 AM, Jean Louis <bugs@gnu.support> wrote:
> 
> ----- Forwarded message from Jean Louis <bugs@gnu.support> -----
> 
> Date: Sun, 6 Oct 2019 08:17:16 +0200
> From: Jean Louis <bugs@gnu.support>
> To: address@hidden
> Cc: hyperbole-users <address@hidden>
> Subject: Re: [Hyperbole-users] Need more usage examples (Was:
>    Subscription within Hyperbole did not work, do I need to
>    repeat sending bugs?)
> 
> * Robert Weiner <address@hidden> [2019-10-05 21:45]:
> 
>> Yes, just set the button highlighting face to have the underlined
>> attribute or any other you want.  When you say you want the buttons
>> to be links, do you mean Emacs pushbuttons or do you just want them
>> to highlight when the mouse rolls over.  All of this is possible, we
>> just need to understand how it will help beyond the way things are
>> now.
> 
> In my opinion, if Hyperbole is recognizing those implicit links in the
> MANIFEST file or other similar files, it should underline those links
> automatically. That could be friendlier for user.

Implicit buttons are not recognized until activated since they can take an 
infinite number of forms and the computational cost is too great.  Explicit 
buttons could be underlined but we prefer color and the easily recognizable 
delimiters.  We will have a look at the visual representations of links in a 
popular web browser for comparison and consider this.

> At this point I have
> use for MANIFEST files, for large collection of PDFs, HTML files and
> similar. Better said, underlining shall be automatic and not left to
> users to set it up, just a thought.

Once you know MANIFEST entries act as links, why is there any issue?

> 
>>>  Additional issue is that duplicated Hyperboly buttons/links are
>>> not recognized until the file is saved. I hope you known that.
>> If a button exists in a saved file and then it is copied within the
>> same file, the copy should work just as the original.
> 
> It is not that I am saying.

You use of duplicated is what was confusing.

> I have tried it today with Hyperbole 7.0.6
> and it works like this:
> 
> 1. I am opening a file like `newtext'
> 
> 2. Then I do C-h h E C and I add for example https link.
> 
> 3. The button appears like: <(Some Example)>
> 
> 4. Pressing M-RETURN on that link does not work. In my opinion this is
>   a bug. If I have added the link the link should work right away.
> 
> 5. I need to save the file `newtext' and after saving the link begins
>   to function.

Yes, this is a valid point.  In the original design, we expected users would 
create many explicit buttons, save the file and then use them.  But with the 
speed of file saving now, maybe we should change this and make it more 
interactive.

> 
> When creating links that are not to work immediately, maybe Hyperbole
> shall ask user if the file has to be saved, or maybe it shall enable
> the link right away without saving it.

We should be able to fix this.

> 
>> Did you have trouble figuring out how to use the menus to create a
>> button or to activate one with the Action Key?
> 
> Yes, I had serious troubles for multiple times, until I finally
> figured it out. Not that I am beginner.

Try to explain your issues other than the one above so everyone can understand.

> Then I figured out how to use menus, and I rather use key
> bindings. And activating is fine too. Thanks.
> 
> There is this issue with terminology.
> 
> For example "Global Button" is something I would expect to appear in
> text files as a button or link that works everywhere in every
> buffer. So far I understand, it does not work in every buffer, it is
> just a menu option.

Correct.  We call it global because it can be activated regardless of the 
buffers onscreen.  Your issue is that activating something by name from a menu 
is not what you would call a button.  But this has been this way for many years 
and yours is the first issue we have heard on this.  I like the idea of 
wiki-like buttons that could appear in any buffer and will think further about 
that.
> 
> That is incorrect terminology if it is only a button in the menu, and not a 
> button that one can insert in the text. Having possibility to
> insert a true global button anywhere in the text files just by using
> its label or name would be usable. I have no special use of menu
> option created under Hyperbole menu.
> 
> "Global Button" is term that is not aligned with "buttons" in
> Hyperbole.
> 
>> Did you have trouble understanding how the HyRolo contact manager
>> works?  Probably not.  If anyone can list specific user-level things
>> that are confusing or not well documented, we will work to eliminate
>> the confusion.
> 
> That one is not quite usable for me. I did try using it, but I run
> into trouble, Emacs was hanging on search within HyRolo, and I have
> filed those bugs, you cannot replicate it if I remember well. It looks
> as plain text contact management.
> 
> I have my own contact management which is based on the SQL database,
> so currently I have no use of HyRolo. That could be used just as
> temporary solution. It has no structure for structured fields of
> contacts, so it cannot be exported, exchanged easily, for me I have no
> use with it. Further, my contact database has too many entries that I
> think it would not be feasibly for simple text search.”

So, you have your own solution for this and don’t need to worry about it.  We 
have logged your issues though.
> 
>> Your issues were likely thinking about ways you can integrate
>> Hyperbole into your highly personalized workflow, which in your case
>> is certainly a fairly complex problem.  Or maybe in trying to
>> understand the myriad capabilities that Hyperbole allows (like
>> trying to understand all of core Emacs or professional-level
>> programming).  We will keep adding useful features that will help
>> you but don't think that getting all you want in exactly the form
>> you want will ever be a simple or fast process.  On the other hand,
>> I think it will be a useful and educational one.
> 
> You are very kind, thank you. Yet do not just think for me, just think
> for overall fundamental usability of Hyperbole for users. No need for
> me for some special options.

Ok.

> 
> In the meantime I have programmed global indexing system HyperScope,
> and I am using it with many indexed entries, now it is over 15000
> entries. It is like data tree based hyperlink index to anything, PDF
> page numbers, or files, Internet pages, videos, everything
> works and it uses PostgreSQL database. One PostgreSQL data type is
> array, so I could use this data type to provide link to the file, and
> to define the links inside that could involve certain words and link
> those words automatically. That could be a feature similar like in
> Hyperbole, just without the file in the directory, it would be stored
> in the database related to the same file.

Yes, we will likely add this capability.
> 
> From Hyperbole I still have expectation to work similar to
> orglink-mode, https://github.com/tarsius/orglink to have links in any
> buffer. But orglink-mode does not work for me, there is some error.

Ok.

> 
>>> They are good, but for advance users. As I have demonstrated, Org
>>> links have their settings, and I think Hyperbole could use the
>>> ready code from Org links to make it available in every other
>>> mode. Isn't it?
> 
> Alright, I guess that will be in upcoming release.
> 
>>>> In the forthcoming release, all of Org mode's link types are
>>> also supported > by the Hyperbole Action Key, so you'll be able to
>>> activate them as well in > Org major or minor modes.  We could
>>> make them work in other modes but > haven't thought much about
>>> that yet.
> 
> Links shall work in every mode, or at least should have possibility to
> be activated in every mode.
> 
>>>> I have built indexed document repositories that can be
>>>> referenced easily within email or other text.  See
>>>> hib-doc-id.el.
> 
> Why don't you make doc-id-indices customizable in Hyperbole group?
> 
> I have tried to understand it. If I get it right, I am supposed to
> define first the index entries, yet example is not sufficient to start
> with it.

This has not been touched in a long time.  We will have to look at it.

Thanks for the feedback.

Bob