2014-02-11 18:10 GMT+08:00 Edward K. Ream <edream...@gmail.com>:
> This post attempts to explain, as clearly as possible, the strengths and
> weaknesses of @file, @shadow and @auto
>
> ===== @file: the platinum standard
>
> @file will always be the best alternative in situations in which it can be
> used.  The reason is simple: it preserves all outline data (including clone
> links) and fully supports all of Leo's features, including section
> references.  Last, but hardly least, Leo can read and write external files
> created with @file more than twice as fast as files created with @shadow and
> @auto.
>
> Otoh, @file inserts sentinels into external files, so it will often not be
> possible to use @file in collaborative environments.
>
> ===== @shadow: often best in collaborative environments
>
> Like @file, @shadow supports all of Leo's features, including section
> references.  Unlike @file, @shadow inserts no sentinels in the public
> external files.
>
> @shadow will work very well for **slowly changing** files, files that are
> infrequently changed outside Leo.  When changes *are* made outside Leo,
> @shadow incorporates the changes in the *existing* outline structure.
> @shadow can *never* create new nodes:  all changes are "shoehorned" into the
> existing outline.  This may or may not be convenient, depending on the
> situation
>
> ===== @auto: a gateway to @file or @auto
>
> When reading an @auto file, Leo creates outline structure from the external
> file itself.  The new @auto code only adds organizer nodes to the imported
> structure.  Furthermore, significant changes to files structure will prevent
> the new @auto code from adding organizer nodes.   Otoh, we expect that
> significant changes to outline structure will be unlikely if those changes
> are made outside Leo.
>
> The new @auto code could conceivably make @auto useful in more situations,
> but @auto still suffers significant problems.  It is slow, more error prone
> than @file, and does not support section references. @auto will most
> typically be useful in situations:
>
> 1.  As a way of studying other people's files.
> 2.  As the first step in creating @shadow or @file nodes.
>
> I often combine these by converting @auto trees to @file trees for study.
>
> Alas, the new @auto code has not fundamentally changed how @auto will
> typically be used.  Perhaps the most important feature of the new @auto code
> is that it allows clone links and (not-yet-implemented) uA's to be added to
> @auto trees.
>
> ===== Summary
>
> @auto is slower and more error prone than @file, and @auto does not support
> section references.  As a result, @auto will never replace @file for Leo's
> sources.
>
> The new @auto code does promise some additional benefits, but many people
> will likely prefer @shadow in collaborative environments.
>

thanx for compared them so detail;

but through words, look like @auto just base bad name?
flow Zen of Python
- simple is better
- every Leo action is simple and do only one things is better
- so
    - @guess foucs: As a way of studying other people's files.
    - @prepare foucs: As the first step in creating @shadow or @file nodes.

in fact i gree so much golang's spirit:

    if one futuare can not
    significantly improve the development efficiency
    ingore it!

as disussed, the Leo big problem is :

CAN NOT EASY EMBDED NORMAL TEAM DEVELOPING FLOW

so the @auto is seem not clear and significantly future?

> Edward
>


-- 
人生苦短, Pythonic! 冗余不做,日子甭过!备份不做,十恶不赦!
KM keep growing environment culture which promoting organization be learnning!
俺: http://zoomquiet.io
许: http://creativecommons.org/licenses/by-sa/2.5/cn/

-- 
You received this message because you are subscribed to the Google Groups 
"leo-editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to leo-editor+unsubscr...@googlegroups.com.
To post to this group, send email to leo-editor@googlegroups.com.
Visit this group at http://groups.google.com/group/leo-editor.
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to