Can't assign a note to multiple shapes

[Paolo F Cantoni]

One can apply a note (or contraint) to multiple lines at one time.  However, you can't do it for multiple shapes only (or a mixtrue of shapes and lines).

Please fix...

A reverse example of the lines vs shapes dichotomy in EA.

Paolo

[KP]

On the UML Elements toolbar next to the Hyperlink icon there is a Note Link icon. Simply click and drag!

HTH,

Neil

[Paolo F Cantoni]

Thanks Neil....

However, I observe you can't link a note attached to a shape also to a line...

Paolo

[thomaskilian]

You have to do it vice versa. First attach to line, then to the element. Not orthogonal but see your tag

[Paolo F Cantoni]

At least we agree orthogonality is good!

FWIW, my tag line is an contraction of an aphorism I've used for the past 15 years....

Some systems (precious few) are consistently correct.  You can "bet your life on them".

Many systems are consistently incorrect.  But that's sort of OK, because at least you can consistently adjust for their incorrectness.

The real bitch systems are those that are inconsistently incorrect.  You basically can't trust them.

[glb]There is no such thing as an inconsistently correct system.[/glb]  (Asserts Paolo Cantoni).  Therefore, aim for consistency in the hope of getting correctness.  Because being consistently wrong is MUCH better than being inconsistently wrong.

YMMV - works for me!  

BTW: Consistency is much easier to monitor than correctness...

Paolo

[thomaskilian]

That reminds me on a discussion about Gödel a short time ago.

Is there any famous consistently correct system you can name?

[Paolo F Cantoni]

The example I used to cite when giving courses was given to me by someone else, so I haven't actually seen this.

The story goes....

When the Porsche Tiptronic first came out, the user manual was reported to have said words to the effect that:  "On a very good day, under exceptional circumstances, it is theoretically possible for a human driver to make better choices than the tiptronic".

I was told this by a porsche driver...  I've never seen the manual myself...

I know it doesn't quite fit the bill, but I think it's a good story...  Certainly, usually raises a laugh...

In any event, the point of my aphorism is that ANY inconsistency is, by definition, a defect!  Many (probably most) developers see rectification of inconistencies as enhancements.  In my view they've missed the point...

Paolo

[thomaskilian]

Indeed a funny story But you really should read what Gödel told us: There can't be any complete axiomatic system. If it is complete, it will contain an axiom saying "What I'm telling is acutally wrong". Having proofed that, Gödel led Russell's Principia Mathematica ad absurdum. You can take it for real systems: If something proofed to be too good, something is wrong.

[Paolo F Cantoni]

I didn't say it was complete, I just said it was self-consisent.  (Completely consistent does NOT imply complete - Thinking 101  )

I understand (I think) Gödel's point and I'm not going to argue with it...  I just don't think it applies to my argument.

Paolo

[sargasso]

Is a pseudo-random number generator inconsistently correct?

:-/

bruce

[Paolo F Cantoni]

Only if it generates pseudo-random numbers...  

If it generates truly random numbers then it's not correct...  However, I do suspect it would be consistently incorrect under my defintions...

Getting serious for a moment, many pseudo-random number generators have the ability to generate a repeatable set of pseudo-random numbers for testing purposes, whether that makes it consistently correct is a moot point.

Paolo