Fixing broken links

[rikona]

Hello Ralf,

Wednesday, February 17, 2016, 3:47:26 AM, Ralf wrote:

> On Tue, 16 Feb 2016 22:01:32 -0800, rikona wrote:
>>Is this really as hard as it sounds? :-((

> My knowledge about bash/dash is limited to things I need, so I might
> miss some useful commands to handle pitfalls, I would experience
> when writing a script to solve your issue. For me it would be hard
> and perhaps impossible.

Not good news for us non-programmers...

> It doesn't help you with your problem now, but for the future
> consider not to use that much likely unneeded links or to write a
> wrapper for the (ln) link command, that helps you to manage links
> already when you create them (by at least generating a file with
> comments to the links), assumed you're using ln, when creating
> links.

I usually create the links by a drag/drop in konq, telling konq to
make a link. This is very easy to do, but would complicate the
'wrapper' idea, which is a good one. I almost never use ln.

> In computer programming languages jump to/go to commands, that do
> not jump to a sub-route are frowned upon. Only Assembler requires to
> use some kind of branch commands. For high-level languages there might
> be a few exceptions for e.g. indexed jumps by pointers too. However, in
> general jumps tend to cause so called "spaghetti code". When
> programming you at least could add comments, explaining the knot garden
> a little bit.
> I handle soft-links in a similar way as I handled jumps, when I wrote
> programs (this was around > 20 years ago). The more careless we use
> soft-links, the more likely we will lose control and often a link might
> be unneeded and should be avoided.

I use links a bit like tags in HTML. For example, there may be info
about topics B, D, and H, that also is important and relevant in topic
A. I put the relevant links to B, D, and H in the topic A dir. I find
this a VERY useful way to organize the complex, interrelated
'knowledge' on the computer.

It sounds like one would have to build a data base of links using CLI
commands, then process the data to document the links and, especially,
to find multi-link loops.

I'd consider other ways to do this if anyone has suggestions...

Thanks for the support.

-- 

 rikona