Hi Dale,
 
I don't think I'm in favour of link order being shown in the workspace.  I 
really don't think that users should have to care about the order that the 
links are drawn, and exposing this will just increase complexity in a tool that 
is currently very easy to use.  At least, once you manage to learn to "Think 
FME".
 
If there is a setting that improves performance (such as clippers or bases 
first) Workbench should automatically refactor the workspace to achieve that 
effect.  If there are cases where a user might want to override that setting, 
it should be through an option in the transformer, not by changing the order 
they draw links or add datasets to the workspace.
 
Apart from being fragile and unobvious, a problem with ordering the datasets is 
that the further away from the source I get, the less sure I am that this order 
will be maintained.  In fact, after several years of using FME I'm still not 
all that clear on how features move through the workspace, and it bothers me a 
bit that some of the really cool things that Mark and others have been able to 
come up with rely on this knowledge.
 
FME has come a long way in making data translation pain-free, but there is 
still a bit of the "high priests of GIS" aura around understanding how it 
really works.  GIS is rapidly moving into mainstream IT, and I believe that 
fostering practices which encourage this mystique is a mistake. It is tempting 
to just expose the complexity and allow users to deal with it, but in the end 
of the day I want a tool that allows me to do my job quickly and efficiently, 
without having to remember minutiae when I come back to it after a couple 
months of focussing on something else.
 
Jason  (long since assimilated)  
 
By  the way, my co-workers were giving me a hard time last week because I was 
being more argumentative than usual.  Apologies if that has come through here 
as well.  :)

________________________________

From: Dale Lutz [EMAIL PROTECTED]
Sent: Sat 2006-09-16 10:17 AM
To: [email protected]
Subject: RE: Order of features read -- was SpatialFilter Usage



Hello Jason, others,

 

This was a fascinating thread, filled with hockey references (and training camp 
is only just starting up this week!).

 

But the root of it I think is a bit unfortunate.  From what I can tell the 
contentious issue was about the order of features flowing into things, and the 
fact that if you're doing clipping or spatialfiltering, you're MUCH BETTER off 
if somehow you can get the clipper/filter features to arrive first than they 
straggle through the door at some point mid translation.

 

So, a few comments:

 

1)    We should emphasize this fact in the documentation for the Clipper and 
SpatialFilter - by way of this email I'm asking that this be done

2)    We should then also give a description of how you can control the feature 
order in workbench - there actually is an obvious and blatant way of doing it, 
but folks may not know.  I'll describe it below.

3)    In general, the order that links are drawn shouldn't be something we'd 
want people to care about.  That is a bit subtle, but I am willing to entertain 
ideas on how we could make that more explicit if people think it is useful to 
know/control. For example, we could have a mode where each link gets a tiny 
annotation number (1,2,3 etc) as it emerges from a port to tell you its order 
in how the feature copies will shoot out.  I think we have technology to do 
this, probably it would be an advanced option, because again, I mostly think 
people shouldn't have to worry about that expect in rare cases. (I'm interested 
in feedback on this idea, perhaps in its own thread here on the list, or in 
person at the UC next week).

 

RE: #2, the recommended way to get the clippers/filters into factories first is 
just to set up a reader for them, and make that reader the first reader.  You 
can control the order that the features are read - they are read from each 
reader in order from top to bottom as they are listed in the navigation pane, 
and indeed, the writers are written to in the order they are listed in the 
navigation pane.  You can move readers/writers up and down as need be in the 
navigation pane not by dragging unfortunately, but by right clicking on them 
and saying move up or down.

 

If you are reading a database to get both the clippers and clippees, then use 2 
readers on the same database, one choosing the clipper table, and the second 
one choosing the clippee table.  Then the order of things is explicit and not 
surprising.

 

Should we number the readers and writers in the navigation pane to drive home 
this point?

 

All feedback welcomed as we, like the Borg, strive for perfection.

 

 

 

 

Dale

 

----------------------------------------------------------------------
Dale Lutz              Safe Software Inc.                 [EMAIL PROTECTED] 
<mailto:[EMAIL PROTECTED]> 
VP Development         Surrey, BC, CANADA        phone: (604) 501-9985
                       http://www.safe.com <http://www.safe.com>          fax: 
(604) 501-9965
----------------------------------------------------------------------

 



Join us at the FME Worldwide User Conference Sept. 21-22, 2006 Vancouver BC 
Canada. For more information, visit www.safe.com/2006uc.   
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/fme/

<*> Your email settings:
    Individual Email | Traditional

<*> To change settings online go to:
    http://groups.yahoo.com/group/fme/join
    (Yahoo! ID required)

<*> To change settings via email:
    mailto:[EMAIL PROTECTED] 
    mailto:[EMAIL PROTECTED]

<*> To unsubscribe from this group, send an email to:
    [EMAIL PROTECTED]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 


Reply via email to