RE: CamelCase For Procedures Names

From: <>
Date: Tue, 22 May 2012 08:27:50 -0400
Message-ID: <>

I believe you are right, and Steve does use plural for his tables. But I also believe it is redundant to do so, and unnecessary, and therefore I like Michaels response :), (but Michael, I still like the underscore... :)). I don't really have enough vested to consider this a religious war (yet), but I just happen to be working (in my spare time) on creating standards documents, and the code to check the standards, so I am kind of interested in the dialogue. (Also how some of the stuff gets created, ie indexes and not null check constraints for instance...). Therefore, I can be persuaded given reasons I have not thought of yet. There does seem to be a lot of variation when it comes to 'convention'. It is like an art. For the most part we agree in principle, yet no two shops name everything exactly the same way. I actually find myself surprised a bit - after all these years.

As far as pl/sql goes, we don't do a lot with that so I'm letting that slide -- in that Steves standards are fine.

Triggers are interesting. How about this convention? TABLE_XYZZZ (X¾fore After, Y=Row or Statement, ZZZ can be any combination of Insert, Update, and Delete)

Joel Patterson
Database Administrator
904 727-2546
From: Jeff Chirco [] Sent: Monday, May 21, 2012 3:39 PM
To: Patterson, Joel; oracle Freelists
Subject: RE: CamelCase For Procedures Names

Really, I feel table names should be plural. I believe Steve Feuerstein recommends this too. It is a Table of EMPLOYEES not a Table of EMPLOYEE.

From: [] Sent: Monday, May 21, 2012 11:44 AM
Cc:;; Jeff Chirco; Subject: RE: CamelCase For Procedures Names

I certainly agree about the prefixes - and the suffices for that matter. I just made an exception for tables - but optional. Something not mentioned, but not optional IMHO is to make table names singular. It is the EMPLOYEE table, not EMPLOYEES.

Joel Patterson
Database Administrator
904 727-2546
From: Michael Moore []<mailto:[]> Sent: Monday, May 21, 2012 11:47 AM
To: Patterson, Joel
Cc:<>;<>;<>;<> Subject: Re: CamelCase For Procedures Names

I think the advantages of having a standard prefix outweigh the disadvantages. One such advantage I found unexpectedly. We start all of out table names with T, like TVENDOR. The big advantage comes in verbal conversation. When, in a design meeting, somebody says 'tee-vendor' , it is immediately clear that they are talking about the database table, and not a person that sells things. Of course you could say, 'the vendor table', but when the names of tables are mentioned thousands of times over the course of a month, the savings in time and effort adds up. Other advantages may be specific to your shop. I.e do you use Perforce? Do you produce reports that contain a mishmash of dictionary object types. One suggestion I would make however is not to use a _ . ie. not T_VENDOR, but TVENDOR. With only 30 characters, every character counts. Sometimes you will want to combine names, for example suppose you want at table that maps vendors to customers. You might want to call it Tcustom  er_vendor_map. Then suppose you want to name an index for this table; tcustomer_vendor_map_ix. In this particular example still did not push beyond the 30 character limit but many table names will be more than 6 to 8 characters to start with. This might seem like a trivial problem, but it can be very annoying when you are trying to follow the index naming standard and somebody has created a table name that's alerady 30 characters long. Use abbreviations where possible. Instead of Tbilling_revenue_stage, use tbil_rev_stg. Maybe the first time somebody looks at it, they won't know what it means, but some poor schmuck like me is going to have to type that table name a thousand times over the next 5 years, so keep it short. IMHO. Mike
On Mon, May 21, 2012 at 6:00 AM, <<>> wrote: I like the example, it pretty much says it all (43 chars though).

What about suffixes or prefixes like R_for_procedures (or _R), or P_for packages, I did not catch Steve mentioning Packages and Package declarations in the paper? (btw I don't believe tables need a prefix or suffix).

I do like what Steve says about avoiding 'get' with regards to a function.

Joel Patterson
Database Administrator
904 727-2546<tel:904%20727-2546>
-----Original Message-----
From:<> [<>] On Behalf Of Andy Klock Sent: Friday, May 18, 2012 9:11 PM
To:<> Cc:<>; oracle Freelists Subject: Re: CamelCase For Procedures Names

The most important thing is consistency by a defined coding standard that everyone must follow. I don't even think it's because it makes the code more readable (though it does) but more because it saves so much time not having to think about what and how to call stuff. Developers should spend more time on the actual logic than whether or not to use CamelCaseWhichIsAnnoyingAsHellWithLongNames or easy_to_read_names.

An old friend and colleague (who I know is a lurker here:) turned our old team onto a document very much like this:

Make amendments to fit your taste. For me it made all the difference in the world.


> -----Original Message-----
> From:<> [<>]
> On Behalf Of Jeff Chirco
> Sent: Friday, May 18, 2012 5:39 PM
> To: oracle Freelists
> Subject: CamelCase For Procedures Names
> I am settling a debate about allowing CamelCase procedure names within a
> package.


Received on Tue May 22 2012 - 07:27:50 CDT

Original text of this message