PowerDesigner vs Sparx Enterprise Architect
enigma76
sparx enterprise architect for the lowest price. I believe that power
designer does a lot more, but there is a comparison between the two
products? Sybase has made the comparisons?
Let me know what are the differences and what to lose by using
enterprise architect. Sparx EA runs as the object and data modeling?
Has anyone used both products? Do you have experience on?
Are there any articles that compare this type of product, eg
PowerDesigner, Sparx EA, Artisan Software, IBM Rational, etc?
Thanks for the info
gtss
You will be losing
- conceptual and logical data modeling
- extended mappings to other types of models
- the possibility to define data mappings for migrations/conversions
(in the ILM)
- the very extensive customization capabilities that are built in PD
- other features aimed at enterprise architecture
Essentially, you will only be doing data modeling at the physical
level using an UML profile, i.e. designing classes that are
stereotyped as tables.
I have tried Sparx EA myself for data modeling. All I can say is that
it felt somewhat unwelcoming - most likely because I have been used to
Oracle Designer and ERwin, where designing relationships, keys and
constraints are straightforward. In EA these did not felt as
straightforward as I would have liked.
Having said the above, the C# developers around me seem perfectly
comfortable with importing my physical model into EA and examining it
in this tool.
If most of what your team is doing is OO design and development, with
a small enough data modeling effort accompanying the OO design, EA may
be a "just barely good enough" solution for data modeling.
There are comparison documents out there between PD, ERwin and
ERStudio, however I am not yet aware of any that compares PD and EA.
Are there any other factors besides cost (yes, we cringed at the PD's
annual support costs too!) that may driving your attempt to switch ?
Regards,
Gabriel Tanase
sandr...@tec.govt.nz
modelling and EA for modelling our physical environment, applications
and enterprise architecture. This plays to the strengths of both
tools.
EA is poor at data modelling, PowerDesigner is poor at everything
else it really depends on what you want out of your modelling tool.
Both have their quirks and take a lot of work to master, both are
poorly supported and have barely helpful manuals and help text.
Mastering the art of reporting out of either requires aspirin and
alcohol. Both are buggy and it is a chore to get them fixed
A significant difference between the two is that EA allows multiple
users to maintain the same model at the same time but this is a two
edged sword as it doesn't handle versioning well, allows creation of
duplicate objects and your models can end up in a mess.
If you use the repository PowerDesigner operates a check in/check out
process like source control and handles merging changes made in two
different copies of the same model, how well is dependant on the skill
of the user.
It would be useful to know what you would want to do with a modelling
tool to give reasonable feedback.
I am more experienced with PD, having used it since 1998, I have only
been using EA for 3 years. I still swear at both of them, but hey one
of the hidden features of all software is to drive you crazy
Regards
Sandra
Mark Brady
> EA is poor at data modelling, PowerDesigner is poor at everything
> else it really depends on what you want out of your modelling tool.
Poor at everything else? Can you provide an example of this?
> Both have their quirks and take a lot of work to master, both are
> poorly supported and have barely helpful manuals and help text.
Poorly supported? You mean there's not a lot of resources available or
do you mean when you contact support you don't get help. I've found PD
to have one of the best help systems for a majority of features. Not
so much with the VB and not at all for the GTL.
> Mastering the art of reporting out of either requires aspirin and
> alcohol.
Yes, the report writer in PD is a pain. I believe that they are
working on a better interface for the future.
>Both are buggy and it is a chore to get them fixed
I run into bugs regularly with PD, but I've seen many of them fixed in
releases. I can point to patches and say, that's fixed because of my
bug report. I've never had that happen with another product.
> A significant difference between the two is that EA allows multiple
> users to maintain the same model at the same time but this is a two
> edged sword as it doesn't handle versioning well, allows creation of
> duplicate objects and your models can end up in a mess.
You mean editing the same FILE, not the same Model. In PD you can
surely work on the same model as someone else.
> If you use the repository PowerDesigner operates a check in/check out
> process like source control and handles merging changes made in two
> different copies of the same model, how well is dependant on the skill
> of the user.
This is true of every Version Control System in the world. Merge is an
art not a wizard you can click through.
gtss
1) Help: general PD help good or very good, VBscripting (and
customization) not so good, GTL poor. Also, the metamodel
documentation is not very good; definitons of class properties are not
clear some/many times. In fairness to Sybase, they send you to the MS
online docs for general VBscript programming documentation, they don't
intend to include that in PD's help.
Yet, the customization and scripting possibilities in PD seem to me as
way ahead any other major data modeling tool, especially if you're not
a (Java) programmer (i.e as opposed to Java Eclipse-based modeling
tools, such as IBM InfoSphere Data Architect, in which you can code
and add plug-ins to your heart's content, however that's not really a
job for a data modeler, since requires you to know Eclipse EMF -
harder job than understanding PD's metamodel!)
2) Merge: "Merge is an art not a wizard you can click through."
I agree in principle, however I'd like to say that I found the PD
Model Merge the most reliable and most intuitive one in respect of
what it chooses to do or not to do by default when you say "Preserve
Modifications". Practically everything I had done in downstream models
(LDM and PDM) is preserved and not overwritten by the Merge's default
behaviour after a re-generation from an upstream model (CMD and LDM
respectively) with "Preserve modifications". For my way of working it
is already almost wizard-like, I trust it with every simple change
that I make. For bigger stuff I check in more detail anyway, since it
usually highlights some stuff I may have forgotten to do. YMMV.
On Jan 13, 5:57 pm, Mark Brady <fodb...@gmail.com> wrote:
> > Both have their quirks and take a lot of work to master, both are
> > poorly supported and have barely helpful manuals and help text.
>
> Poorly supported? You mean there's not a lot of resources available or
> do you mean when you contact support you don't get help. I've found PD
> to have one of the best help systems for a majority of features. Not
> so much with the VB and not at all for the GTL.
>
Mark Brady
Yep. the place where the metamodel help falls down the most is in
describing HOW to use certain properties. I tried to add an object to
a collection but got an error. I had to do something else to be able
to add that object to the collection. NOWHERE was that procedure
documented, nor is that particularly good programming. If an object
has a collection of objects, then I should be able to add an object of
that type to the collection. If an object isn't "well formed" until
certain conditions are met, those should be clearly documented.
dani...@itcentralstation.com
Users interested in these solutions also read reviews for MEGA HOPEX. This user writes that MEGA is, "becoming the authoritative source of current state application architecture." You can read the rest of his review here: https://goo.gl/4WbQ2C.