Hello,

I would like to get a feel for how people are naming their reports, HTML forms and other files in the WebFOCUS (Developer Studio) environment. We are currently using an IBM-style eight-character file name convention that is not very well-received by most of the people involved. Since our environment is not hosted on a mainframe, we have the ability to create file names longer or shorter than 8 characters, but are having trouble deciding on a convention.

One thought was to use Bumpy Caps or Camel Case, in which each distinct word in the file name is capitalized, i.e. TimeAnalysisByCostCenter.fex The only problem here is that WebFOCUS allows you to mix case when creating a new procedure, but then reverts the file name to all lowercase once the procedure is created. Has anyone successfully named .fex files in mixed case? If so, how?

Thanks for any insight you can provide regarding your preferred naming convention, the above 'mixed-case' question, or life in general.


sleepy

This message has been edited. Last edited by: sleepyEDB,

You do not say where the issue is.

In MRE 76x any name you like (no special characters) up to 64 characters, any case.
Because the users do not see the real name.

WebFocus its' self does what it does to try and make the platforms uniform, MS is not case sensitive Unix is.

Best convention is no spaces, use under-score (x_y), use single case and it's platform agnostic.

Just be greatful that you are not on a release that only allows 8 characters.

Make a choice and impose it.

quote:

Originally posted by JG:
You do not say where the issue is.

You are correct...my apologies. The issue was in the Report Development -> Data Servers -> EDASERVE environment.

quote:

Originally posted by JG:
In MRE 76x any name you like (no special characters) up to 64 characters, any case.
Because the users do not see the real name.

You are correct again, and I should have done a bit more research before posting. Our users have been encouraged to develop within the MRE for other reasons, and now that I know the MRE allows Bumpy Case, this solves my problem...thanks!

quote:

Originally posted by JG:
WebFocus its' self does what it does to try and make the platforms uniform, MS is not case sensitive Unix is.

I see. The fact that WebFOCUS forced the file name to all lower case makes sense now.

quote:

Originally posted by JG:
Best convention is no spaces, use under-score (x_y), use single case and it's platform agnostic.

Thanks. While I don't think there's any chance of us moving the development environment to another platform, I'll keep that in mind.

quote:

Originally posted by JG:
Just be greatful that you are not on a release that only allows 8 characters.

I am.

quote:

Originally posted by JG:
Make a choice and impose it.

We will!


I have marked this thread as SOLVED, but I'm still interested to see what naming conventions people are using and maybe an example just out of curiosity.


sleepy

This message has been edited. Last edited by: sleepyEDB,

I always try to name it with underscore's, the first part is the group (gl= global includes, rd = some reportgroup, st = style, etc)_ class _ name.

When you have a lot of fexes, this is easy to browse withing your fexes using a keyboard on the list in dev studio.

I'm partial to "CamelCase" to avoid spaces, especially since WebFOCUS does not allow spaces, and don't care for under_score that looks so "nineties" (or even earlier).

In WebFOCUS I've started using hypens since CamelCase is turned into-all-lower-case and hard to tell what meaning the filename may have been given.

However, I've noticed that MAINTAIN does not like hyphens and causes errors when EXECuting reports into HTML stacks.
Has anyone else noticed this ?

Thanks

We name our fexes in two steps.
First we create an 8 character codename, which holds 6 numbers and two characters, the number is the actual date when we create the fex, the first character is the name of the programmer and the second is a sequence for that day.
So to day my code would be '081129FA'.
Now this name is also the internal name.
After this first coding, we rename or ad a more descriptive name.
In the FEX we start with a command lime
-TYPE &FOCFEXNAME
This shows us that code.
In the footing of the report we do the same so this code is also visible in the report. When user have a problem they can send us this code.
We have an internal process that gives us all the fexes with internal name, descriptive name, directory etc.

So, your program's 'code name' is simply a key or pointer, and used as a lookup to find out who wrote it and what the program does ?

I take it you never write more than one code segment in the same day ?

The decoding 'process' you mentioned is a program that others can run to find the salient details about your code, that prints only the top 'comment' lines out of your code ?

Interesting ...

Charles, my letter code is only the F, the next letter A is the first fex I write on this day. The next would be '081129FB' and so on. (after Z I would go on with 0, 1, 2....but although I'm rather quick, more than 6 or 7 we never did)

This coding system is also used inside the fex. If we need to create a temporary hold file we use more or less the same codes.

I'm a proponent of naming a family of elements, say, launch pages, framesets, dropdown fexes, report fexes, includable fexes all with a prefix common to a given application;
say, your application in an MRE domain/reportgroup/folder is a balance sheet report..
balsht_frameset.htm
balsht_launchpage1.htm
balsht_ddlist1.htm
balsht_ddlist2.htm
balsht_ddlist3.htm etc
balsht_report1_v1.fex ... use v1 for version1, allowing for more versions to be developed
balsht_join.fex this is a common includable
balsht_defines.fex this is an includable
balsht_report1_dd1.fex ... the first drilldown from report1..
you get the idea, yes?
Each application has its own folder inside a reporting group in an MRE domain.
The ONLY file that gets renamed in its Properties box to something pretty, would be the frameset, the only thing to show in Dashboard.
so, the displayname property of balsht_frameset.htm might be Balance Sheet Report
That's all the user sees. all the other elements have unchecked 'show on users list';
Elements on the 'back end', eg in the apps directory, don't show up on users lists in dashboard.
-s

We are relative new to WF and also trying to come up with a convention. I wish the Data Server and Web App have same capabilities like MRE that keeps the short name for the objects and rename it after to give a descriptive name. I think WF should have object-description field along with any objects. The names of the master files brought over from other platforms are very hard to remember which is which.
Anyway, we are kind of using the old way to name any application:
start with 2-letter domain name,
a 4-digit code that identifies any application
optionally letter(a-z) at the end to identify each related objects
In MRE, we will append the description to each objects.
eg. Invoice Inquiry in Accounts Payable domain would like this:
ap0001.hml - AP Invoice Inquiry
ap0001.fex - AP Invoice Inquiry(main)
ap0001a.fex - drill-down to PO details
ap0001b.fex - drill-down to Receiver details

But it is very hard to maintain the consistency because we tend to give a descriptive name right from the start, or when copying, it alway gives a long name or WF's own 8-character code for the new object.

FrankDutch

OK, I guess 26 programs per day is sufficient ?
Although that would be shared by Frank, Fred, and Frida, that's still 8.3 per day each.

Do you have software that assigns the names in sequence, or is it handled manually ?

How do you tell if F?081201 was written by Frank, Fred, or Frida, and how do they know which letters are already used by other F's ?

My brother recently worked for a firm that used what appeared to be completeely random filenames, and we were not able to find any "code," but maybe there was some "method to their pseudorandomness" after all. (We never decoded it if there was !)

We're trying to find a way to name files that will work across platforms, so thanks for the insight.

Long filenames alleviate most of the problems with short names, but a system of some kind is necessary, and Susannah's 'folder-izing' is one nice way to accomplish some sense of order !

Thanks

Charles

Our company is not that big.
We are with two programmers and we agreed about the used letter. I had the first chose since I work for this firm for 25 years....
It's all a matter of communication. We decided it to do it this way. I have not seen a better way. When we started the 8 character boundary existed. Everybody in the company understands the system and the most imported reason is that we are happy with it.

It is not an "official standard" at our place, but we too prefer putting each application related code members in distinct folders, with an incrementing number in the front. This is especially handy to understand the flow sequence in folders with many members.

01_Launch_xxxxx
02_Rpt_Driver_xxx
03_Rpt_1_xxx
04_Rpt_2_xxxx
91_include_mem1
92_include_mem2

In mixed-bag folders, we prefix related components with the same number and an alphabet sequence for internal re-ordering.

1A_App1_launch
1B_App1_report
2A_App2_report,..... etc.

Sandeep Mamidenna