Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - jaimeglz

Pages: 1 2 [3] 4 5 ... 11
General Board / Re: database schema
« on: November 24, 2002, 02:10:04 pm »
If you want to generate SQL for a class, and also generate code --say, C++ or Delphi-- for the same class, this is means is that you are addressing two different artifacts: one would be a table, and the other would be --most probably-- an interface object (say, a dialog).

My suggestion would be to have tables modeled in a database (structural) diagram (t_CA, etc...), and interfaces modeled in an interface diagram (w_CA, etc...). t_CA should be stereotyped as a table.

Hope this helps,


General Board / Re: database schema
« on: November 24, 2002, 01:12:59 pm »
Hi Peter,

You can generate it either for a single class or for a whole package. See "Generate DDL" in help, and keep us posted if you run into any problems. Make sure the classes you want to use for generating SQL are stereotyped as tables, and check that your target RBMS is specified in "Database".

Jaime Gonzalez

General Board / Re: How to build an RTF File
« on: March 06, 2003, 06:59:31 pm »
Hi nir,

The best way to solve this problem is to break-up the use case or process on which the activity diagram is based into several, smaller, use cases or sub-processes.

It is somehow time consuming, but it makes your whole model (not just a particular diagram) much more readable.

The way I divide one use case into several ones, is by means of <<include>> and <<extend>> relationships.

I hope this helps,

Jaime Gonzalez

General Board / Re: Generating DDL
« on: March 01, 2003, 07:20:04 am »
Hi Mohammad,

On 1.: In the menu bar: Reference -> Database Datatypes     click on Add Product  in order to add MySQL. Then, in that same dialog add MySQL datatypes one by one.

But before you do that, let me ask: Are MySQL's datatypes so different from SQL Server's? Testing to see if SQL Server datatypes and code work for MySQL could save a lot of time.

Hope this helps,

Jaime Gonzalez

General Board / Re: Use Case scenarios
« on: February 27, 2003, 07:18:25 pm »
Hi all,

I don't think that the two column approach ("User Action" and "System Response") is a good practice, because use cases can represent manual tasks that have no interaction with a system. For example, in the Human Development Indexes project we are carrying out for the Mazatlan Villa de Flores municipality, we have to describe a "fill family questionnaire" use case: questioner asks family head for family home data detailed in questionnaire (form) page 1; questioner writes family home data on family data on form, page 1...

I agree with Eric et al. that one of the most fundamental aspects of use cases is to obtain the basic path and alternative paths.

I don't write the detailed event list of the basic path or alternative paths in the scenario section of the use case dialog in EA (though I do write a very brief description of these, or at least something that can remind me and/or other modeler or programmer that some alternative path has to be considered as well). But I do obtain a detailed event list from the sequence diagram(s) that describe the use cases' path(s).

However, I do not mean to deny that what is being called "data fill" can be valuable. In order to obtain with EA something very similar to what James is looking for, I would suggest the following:

1. Create a sequence diagram that describes your scenario (prefferably directly under your use case, if this does not interfere with the way you order your project packages).
2. Generate the RTF documentation of the package that contains your sequence diagram (right click on the package -> Documentation -> Rich Text...   ).

The resulting RTF will contain a column formatted event list of your sequence with "Message", "From Object", "To Object" and "Notes". Example:
(message:) double click, (from object:) user, (to object:) ico_system;
display(w_login)... etc

You can actually change the order of the event list columns in the RTF doc: go to the resources tab of your Project Workspace -> expand Templates, right click on RTF Templates, create an RTF template, and modify the sequence of cells in both Sequence Heading and Sequence Line items. When generating your RTF doc, use the "Style" drop-down list to specify the RTF template you have just saved.

I hope this can be of some help.

Jaime Gonzalez

General Board / Re: Outline Numbering
« on: February 07, 2003, 07:06:59 pm »
Hi Erik; hi all,

Maybe this is not the exact solution you are looking for, but here is how I do it:

Tjerk is right: use the "Heading style" drop-down control in the "Rich Text Format Report" dialog. I almost always choose "Max 9 levels - elements are package +1". (For further refinement, create your own RTF template: go to the Resources tab in your project browser, right click on "RTF Style Templates" to create a new template, and help yourself to some fun by editing the RTF templates for the different heading styles. Sparx's documentation says RTF is "unforgiving", but I think that is just an understatement.)

I suggest not to number your use cases: it will always be a real pain to re-number them. Just create the packages in the use case model in the most logical order you can get in a short ammount of time.

I hope this helps.

Jaime Gonzalez

General Board / Re: Issue Priority changes
« on: February 19, 2003, 07:41:26 pm »
Hi again matteo,

With your instructions, I did reproduce the error you mention. Not only that, but found a couple of issues with "issues" as well:

1. If I display the project issues (created through Project -> Issues) by  Alt + 2, these are displated under a "Model issues" tab. I completely agree with you, this is misleading: these should not be named "model issues", since the Project -> Issues dialog clearly says on that these are "Project issues".

2. In the Project -> Issues dialog, there is no way of changing the priority of issues, which are always created with a "2: Medium" priority.

I have to admit that, although I have been using EA heavily during the last ten months or so, I had never displayed the project issues by Alt + 2.

I had always created model issues by dragging the "Issue" icon into a diagram. (By the way, I keep all model issues on an appropriately named package.) I do have to remark that these issues (which I consider as proper "Model issues") never appear together with project issues: they don't come out in the Project -> Issues  dialog, nor in the RTF document created in this dialog, nor in the Alt + 2 display.

So, although these are minor "issues on Issues" (albeit confusing ones, as you correctly point out), we should all thank you for raising them.

If you follow this forum for a minimum a ammount of time, you will see that Sparx is very responsive.


General Board / Re: Issue Priority changes
« on: February 18, 2003, 04:07:19 am »
Hi matteo,

Can you give us a bit more detail on how you created the issue you are mentioning? I could not reproduce the bug: I opened the issue dialog for one of the issues in a project and have had no problem changing the priority. The issue in question was created by dragging the issue icon (in the "Custom" tool box) into a diagram.

Please be aware that project issues (Project -> Issues ) are different from issues that pertain to the model (like the one mentioned in the last paragraph). A project issues would be like the following: "Current project budget underestimates testing time: need to discuss this with sponsor".

A modeling issue, on the other hand, has to do with your subject area. Example: "Accounts Payable is requesting expenses control, but current budget is not ready yet".

Project issues have to do with your process in creating the model, whereas model issues have to do with your customers' or subject area's processes.

Hope this helps,


General Board / Re: Association labels resizing themselves? Huh?!
« on: February 17, 2003, 05:47:10 pm »
Hi all,

Coping with association (and, for that matter, all types of relationships) label spaces resizing themselves is one of the few sources of discontent I have with EA. When I click somewhere in the diagram, get a highlighted association label space, instead of what I was expecting, is oftentimes bothersome.

Thanks for bringing up the point, which is an oportunity to ask Sparx to take a look at this problem.


General Board / Re: postgres support
« on: February 14, 2003, 07:17:28 pm »
Hi Uwe,

I've done a good amount of modeling with EA for a PostGreSQL database. Everything seems to work alright, though I'm not actually doing the SQL code generation, which is the responsibility of the progamming team (that resides in another city).

I'm sending a datatype file to your EMail, so you can import the PostGreSQL datatypes into your model (Reference -> Import Reference Data...   select the file...).

Just in case, please back-up your project before importing the datatypes I'm sending.

Jaime Gonzalez

General Board / Re: How do I update a Data Model?
« on: February 14, 2003, 09:20:18 pm »
Hi Riebert; hi rescobar,

After you have imported your tables into the model, and after you made your diagram tidy: Have you tried importing new tables into another (say, temporary, or staging area) diagram? You can then copy the new table (Ctrl+C) into the tidy diagram (Ctrl+V), put it where you wish, and then you only have to tidy the relationships.

Hope this helps,

Jaime Gonzalez

General Board / Re: Sequence <--> Class diagram relationship
« on: February 04, 2003, 07:17:15 pm »
Hi Raze,

I hope this is what you are looking for:

In your sequence diagram, double click on the message that is intended for the new operation.

In the messsage properties dialog, click on the "New" button: this opens the class' Operations dialog, where you can create the new operation.

Jaime Gonzalez

General Board / Re: Hierarchal system design, package=>class
« on: January 14, 2003, 07:57:04 am »
Hi Per,

I think the problem stems from an inconsistency inherent to stereotyped <<subsystem>> packages.

From a strict UML point of view, a package is just something that groups several elements. It is not a classifier (like a use case, a class, a collaboration or a node), that can have components, attributes, behaviour, inheritance, and so forth.

It is very convenient to pack everything that will be part of a subsystem into a <<subsystem>> package; but how can something that is not a classifier have the kind of behaviour that you are trying to confer to the package?

I had the same problem when I tried to model a complex catalog as a package that contained several tables. The solution I found was to convert the package into a class, and create aggregation relationships between the tables and the class. I even dropped the tables into the class' icon in the project tree, to use the class as the grouping element.

To keep the convenience of using <<subsystem>> packages, a <<subsystem>> class stereotype can be created.

I hope this helps.

Jaime Gonzalez

General Board / Re: Sequence Diagram- branch with previous message
« on: January 22, 2003, 04:19:53 pm »
Hi Dave,

Here's a link to a thread where a plausible solution (proposed by Geoff) was discussed:

Please keep us posted ff embedding your IFs within iteration symbols does not work for you.

Jaime Gonzalez

General Board / Re: Matrix filled automtic?!
« on: January 15, 2003, 08:04:44 am »
Hi Jens,

The relationship matrix is automatically updated every time you create a new element or a new relationship. Check the following:

1. Your source and target packages (upper left corner of the matrix). For example, if a new element is contained outside the package you chose for source, it will not show up with the source elements in the matrix. You can make the root (namespace) of your project both the source and the target, so as to make sure you have everything in your matrix.

2. The type of element (to the right of source and target). If I have "use case" as source and "requirement" as target, the matrix will only show these types of elements.

3. The link type and direction. If I have a dependency link from a use cases and a requirement and I have chosen something different from "Dependency" and "Source -> Target", the relationship will not show up in the matrix.

On diagrams: I do not know of any way to link two diagrams so that changes in one are reflected in the other. However, new relationships made in one diagram between (say) two objects, are displayed automatically in all other diagrams that contain those two objects.

I would recomend for you to create a particular element ("customer" class, "Charlie" object, "deleted" state for "customer" class instances...) only once, and reuse this same element in all you diagrams: just drag the element from the project tree into the new diagram (and, in the case of classes, make sure you do not have "auto instance" on). This way, all changes to a particular element are reflected in all you diagrams.

Hope this helps.


Pages: 1 2 [3] 4 5 ... 11