Robert Bar

Ranch Hand
+ Follow
since Jun 29, 2006
Cows and Likes
Cows
Total received
0
In last 30 days
0
Total given
0
Likes
Total received
0
Received in last 30 days
0
Total given
0
Given in last 30 days
0
Forums and Threads
Scavenger Hunt
expand Ranch Hand Scavenger Hunt
expand Greenhorn Scavenger Hunt

Recent posts by Robert Bar

Originally posted by ken bonzo:

How did you recover the InterruptedException?
do we have to? I do not call the Interrupt method at all

Regards



There are several possibilities to handle this problem: (1) eat exception, (2) throw runtime exception, (3) throw exception with higher level of abstraction, (4) start contigency action etc.

I didn't developed any contingency plans for such exceptions, so I had to consider (1), (2) or (3).


regards,
Robert
15 years ago
1. it looks not good to me. Sorry, I can't tell you more. I don't want to take fun out of developing.

2. see following documents:

- how to write doc comments.

- requirements for writing API specifications

regards,
Robert
15 years ago
Hi,

Generally I agree with Anthony. In most of situations providing javadoc for private members it's a waste of effort.

Mike, see following documents:

- how to write doc comments.

- requirements for writing API specifications

regards,
Robert
[ October 08, 2006: Message edited by: Robert Bar ]
Mike,

- yes, I had the same requirement.

- You met the mentioned requirement, so don't worry. Just document the issue.

regards,
Robert
15 years ago
Hi,

Originally posted by Mike Tilling:

In my implementation, if the user change the working directory,



Can you tell me, when/how does the user change the current working directory? Can I assume it's done before application starts?

Originally posted by Mike Tilling:

in other words :
-Is it required to keep the same suncertify.properties file even when the user changes the working directory?



It's not required to keep the same file, because using Properties API, it is extremely difficult to backup a user's preference data. Just think about the problem of remembering a path to the last current directory. The only thing you can do, is to inform the user about the issue (in help file?).

There is a better solution than using properties API - preferences mechanism. Maybe you can mention about this solution in your choices.txt.

regards,
Robert
15 years ago
Hi,

Originally posted by Rob van Oostveen:
Hi Robert,

Thank you for your reply. So you're saying that I should throw RecordNotFoundException in case this happenes? I don't have much choice though, because I'm stuck with the Interface contract.

If you have another suggestion, please let me know.

Regards,
Rob



No, no, no. Please read my first post again, I suggest to use runtime exception.
As I wrote before, RecordNotFoundException has unambiguous interpretation: it should be thrown if record doesn't exist or is marked as deleted. In any other situation I suggest to utilize subclass of RuntimeException, cause it don't change the required interface, and don't overload the meaning of RecordNotFoundException.

At last, throwing RecordNotFoundException is much more acceptable than eating the exception.

Just don't sweep dirt under the carpet. ))

regards,
Robert
Hi Rob,

Think about consequences of swallowing that exception: lock() returns normally, and client modifies the record without the lock! Another client can lock the same record at the same time.

Generally, you should either complete your contract and return a valid result, or throw an exception. It's a samurai principle : Return victorious, or not at all.

Consider also throwing some kind of runtime exception. RecordNotFoundException signals that record doesn't exist, and maybe it would be better to rest the meaning of that exception unchanged. What do you think about that?

regards,
Robert
Hi,

My client-gui has following structure:

- menu bar (File: Connect, Disconnect, Exit; Edit: Book, Find (All, By Criteria in submenu), Refresh; Help: Contents)
- tool bar (24x24 icons)
- table (ascending alphab. sorting, sorting column header is bold)
- status bar (connection status or last operation status)
- find dialog
- book dialog
- connection dialog (polymorphic)
- progress dialog for long running tasks
- help in separate frame

BTW: can I upload screenshots here?

Regarding to the functionality enhancements:

- I couldn't imagine what kind of framework gui should establish. What does 'framework' mean for them?
- I didn't know how to measure/qualify disruption to the users. What does it mean, when someone says that disruption is minimal? Ultimately, what is 'disruption'?
- Must this disruption be minimal to the clients of the system, or to the users of the library?

Because of this problems, I can't be sure that I've achieved the aim of providing mentioned 'framework'.

regards,
Robert
16 years ago
Hi,

I was determined to finish my assignment before 31 Aug. because of the vourcher's deadline (I started in the middle of July ). Normally I develop Swing applications, so there was no effort spent on learning Swing mechanics and gui guidelines.

BTW: I wrote various testcases for Data class to be more courageous during refactoring. I used Eclipse with checkstyle and findbugs plugins installed.

A big thank you to all who helped me out with their posts and suggestions!

best regards,
Robert
16 years ago
Hi!

General Con: 100 93
Documentation: 70 65
OOD: 30 30
GUI: 40 29
Locking: 80 80
Data Store 40 30
Network Server: 40 40
Total: 400 367

I've worked on URLyBird 1.3.1
My GUI was developed according to L&F Design Guildlines, so I'm a little disappointed with my GUI score.

My project metrics:

lines of code: 4200 (gui:2500, db:750, business logic: 750, network: 200)
number of classes: 55 (gui: 30, db:13, business logic: 9, network: 3)
method lines of code: 2750 (gui: 1500, db: 650, business logic: 500, network: 100).

total effort: 120h, duration: 6 weeks. (documentation: 40h, coding, reading this forum: 80h)

best regards,
Robert
16 years ago
Samuel,

In my interpretation, the statement: "A null value in criteria[n] matches any value" actually doesn't rule out the statement: "Some of non-null values in criteria[n] matches any value". Both statements are logically consistent.

Your interpretation is:

"A null value in criteria[n] matches any value" for that reason "None of non-null values in criteria[n] matches any value".

I think my interpretation doesn't violate the 2nd rule:

"A non-null value in criteria[n] matches corresponding record[n] value that starts with criteria[n]"

and your interpretation violates it (because for you "" is a non-null value that actually matches to nothing).

regards,
Robert
hi Samuel,

are you sure that you take the requirements literally?

Look, the "" value is a non-null value, so it should be treated like any other string, shouldn't it?

regards,
Rob
hi,

my spec says that:

- null matches any value
- non-null value (lets say A) matches any field value that begins with A.

But what about "" string? It is a non-value value that accidentally matches any field value!

Is it possible to have only null value reserved for matching any field value? Maybe "" can be reserved to match only "" field value? Or maybe it is against the spec?

please, help me )

rob