On Apr 13, 1:10 am, wolfgang <
wolfgangl...@gmail.com> wrote:
> Eric is the most productive and conscientious developer, there is none
> equal to him. Morris is writing and rewriting scripts, and so a
> developer too. Therefore they are NOT talking out of the perspective
> of users at all. Though they are certainly most experienced with
> TiddlyWiki - however, they do that out of the perspective as
> developers.
If 'user' means someone who only reads TW documents, then I concur
that I'm not a TW user. However, *anyone* who makes even simple
changes to the content of a TW document can benefit from the
SaveAsPlugin, so it is not just an issue for 'developers' such as
myself.
Take for example, the case of a user who imports a bunch of plugins
and attempts to carefully follow the installation instructions,
modifying various templates, stylesheets, etc. only to discover that,
after completing the necessary save-and-reload, something has gone
terribly wrong and their document has become unusable.
While there is (hopefully) a backup file, it still takes a bit of
'filesystem gymnastics' to remove the currently broken document, and
restore the previous version by locating, copying and renaming the
correct backup file.
> No wonder they need such functions in every case - other than me who
> is not able to do coding and a real and pure user! And I say 'no
> thanks' to 'saveAs' in the core.
With all respect (and I do mean that!), I don't think that you are a
typical user. You have an extensive knowledge of a wide range of
available plugins, and have done considerable research on how to load
those plugins 'on demand', using some advanced techniques such as
bookmarklets and compressed external js files. The typical 'pure
user' has no such knowledge to draw upon.
While I was suggesting that SaveAsPlugin *might* be a good 'core
candidate', as a plugin it can be easily added to a document only if
desired, and even loaded on demand via a bookmarklet (if it is first
converted to an external .js file). Nonetheless, this is not likely
to benefit the vast majority of 'pure users', who generally rely upon
only those features that are included in the standard TW core code
plus, perhaps, a few plugins that may have been pre-installed by some
helpful 'power user'.
> > ... and, in this instance, the phrase "considerable code overhead" is
> > just plain silly, especially when compared with any of the following
> > 'narrow use case' functions currently present in the TW2.3.0 core:
> In these cases I agree with Eric (though not in his use of words out
> of seeming contempt) that functions which are so rarely used and only
> their writer could explain what they do, or even actually are
> dysfunctional, should not be given any space in the core (just opened
> a TW of 2 years ago, my goodness: only 100 kb and everything there
> what one usually and daily needs of a note taking TW!)
My 'contempt' was for the phrase 'considerable code overhead', and
certainly not for the writer of those words (FND). I have nothing but
the highest regard for Fred's contributions to the TiddlyWiki
community and did not intend those words to be taken as any kind of
personal criticism.
> 'Save as' - as it is now in the form of a plugin with 4515 bytes - is
> for me a too narrow use case too. And I wouldn't want it for anything
> in the continually growing standard distribution of a empty TW!
If SaveAsPlugin were to be added to the core code, it wouldn't include
any of the documentation or the slice table that are part of the
plugin tiddler. This reduces the 'overhead' to 2373 bytes.... and, if
only the general "ask for filename" functionality were to migrate to
the core, that code is a mere 1320 bytes.
By comparison, the Crypto functions (which *are* being added back to
the core for TW2.3.1) total 3667 bytes and, while not directly needed
by any TW core code are being included in the core primarily so they
can be used by various plugins without having to either A) include
those functions in each plugin (contributing to even an even more
'bloated' document), or B) creating a dependency on the separate
CryptoFunctionsPlugin (which complicates installation and can easily
result in 'broken' documents if omitted)
In this particular case, I think that having the Crypto functions in
the core is the right thing to do. Similarly, I think that adding at
least the general cross-platform 'ask for filename' function to the
core is very sensible.
In fact, that particular chunk of code is currently replicated in
several TiddlyTools plugins/scripts, including: SaveAsPlugin,
ExportTiddlersPlugin, QuickEditPlugin, SaveFromWebPlugin,
NewDocumentPlugin, ShowLocalDirectory and SaveTiddlerToFile. If this
code were in the core and someone were to install some or all of the
above plugins/scripts, it could actually *reduce* the overall document
size by as much as approximately 8K! (6*1336 bytes)
> kind regards,
as always, same here...
-e