Baseline Show comparison connectors

[Monsieur]

Hello.

I have some "empty" old diags.
I do a baseline on the parent package.

I compare the model with my baseline.


It seems there is a residual Information Flow on the empty package.
I can't find it anywhere nor delete it.

Why this residual is not in the baseline if it's on my model?

[qwerty]

I found that some years ago and stopped using baseline compare. It has been worse before, but finally it still rendered useless for me.

See my last comment on p.1 : http://www.sparxsystems.com/cgi-bin/yabb/YaBB.cgi?num=1330097413/0

q.

[Eve]

First, try a data integrity check.

If that doesn't work, what appears on the diagram in that package? I can imagine the xmi not exporting the connector information (which would show up as it existing in the model only) if you have objects showing on that diagram and connectors between them.

[Monsieur]

I found that some years ago and stopped using baseline compare. It has been worse before, but finally it still rendered useless for me.

See my last comment on p.1 : http://www.sparxsystems.com/cgi-bin/yabb/YaBB.cgi?num=1330097413/0

q.

Simon M wrote on Feb 28th, 2012, 9:37pm:
I've heard of connector styles being different in a baseline comparison, although it's disconcerting, it doesn't appear to be serious.

Not serious, but: This (reported) bug makes the whole baseline compare almost unusable. Running a compare will throw dozens of differences in connectors where there aren't any (it claims the "hidden" attribute to be duplicate). So this is not serious but at least major.

Another reason that makes finding above serious bug harder as it is covered under tons of non-serious fault messages.

q

Q:
What solution do you use then?
Are they other way to compare models (as baseline will normally do)?


Simon M:
Does this will/have been corrected?
Will it be?

First, try a data integrity check.

If that doesn't work, what appears on the diagram in that package? I can imagine the xmi not exporting the connector information (which would show up as it existing in the model only) if you have objects showing on that diagram and connectors between them.

This seems to have suppress the error!  
Where does this error can come from? I want to know the source of the problem and avoid repairing all day long. Is that recurrent?

[qwerty]

It depends on the reason for doing the compare. Depending on that you find different approaches.

q.

PS: Re. Integrity Check: you should run that on a regular (weekly) basis. I don't know why EA doesn't do that like fsck in *NIX.

[Monsieur]

It depends on the reason for doing the compare. Depending on that you find different approaches.

q.

Such as?
Two example of use:
We do them to compare User Requirement between each delivery
We do it to verify that each implementation point was treated and what were all impacts across projects

PS: Re. Integrity Check: you should run that on a regular (weekly) basis. I don't know why EA doesn't do that like fsck in *NIX.

Thanks for the tips, i did not know that!
Is the impact heavy on data (you got a BIG warning when you lunch it)?

[qwerty]

Please don't lunch it since it can harm our health

The integrity check scares you when asking to make changes which seems to imply that some of your work might get lost. That's not the case. Only if you do NOT run it you will encounter bad things happening in your model (as you already experienced).

In order to achieve the comparison: you can use the audit or the a search for modification date. The audit will additionally show who changed something. Knowing that something has changes is usually enough since the persons that have to take care need to know the history anyway (how else could they take care and thrive modeling?).

q.