The Humble PL/SQL Exception (Part 1a) - The Structure of Stored Subprograms

As I said in my previous post, The Humble PL/SQL Exception (Part 1) - The Disappearing RETURN, there are a lot of nuances surrounding exception handling. That post attracted some comments that I thought deserved a followup post rather than just another comment in response.

oraclenerd said (excerpted):

I'm going to have to disagree with you on the internal procedure (in the declaration section) paradigm. What about testing?
...
Now you have 2000 lines in your declaration section and 400 or so in the body. (I've seen it...really, I've seen it). Then you want to change one of those internal procedures...you can't test it without testing the entire thing.

This is to me one of the conundrums with any programming language; PL/SQL just has its own unique aspects.

Any set of procedures can be a testing challenge, since you can think of calling a procedure saying "do this", whereas you can think of calling a function as asking "tell me what would happen if you did this". Sure, it's easier to test a function, because you just have an in-memory return value that you can compare against some expected value, and don't have to worry about changing data by accident or doing rollbacks.

In the situations where I would use the technique of lifting code directly into internal procedures, the code is typically very short and easily verifiable: select count(*) into... followed by an IF test; a sequence of simple assignments that would be cumbersome to turn into one function per assignment; blocks of code that have already been verified in toto by virtue of tests run against the outermost procedure or function; that sort of thing.

After the code is modularized this way:

• The outermost procedure can often be improved and/or optimized simply by reordering the inner procedure calls. For example, to do all the "should I quit early" tests before doing any substantial work. Or to put "set up complicated string value" right next to "use that string value".


• Debugging steps such as removing blocks of code can be performed by commenting out single lines. (Much appreciated if those blocks of code themselves contain multi-line /* */ comment blocks.)


• If a logic problem does turn up in the procedure, it's easier to figure out the part to look at if it has its own descriptive name. And a tool like ctags makes it handy to jump directly to that inner procedure.

The problem with a 2000-line declaration section, to me, is no worse than if you go the package route and must fix syntax errors or track down logic errors within a big package body. I use my favorite SQL*Plus hack to push that code off into separate files once it gets too big.

Anyway, my point is not to advocate using this kind of structure for every procedure or even every big procedure, rather to suggest that if you do restructure a procedure this way, using exceptions instead of return can make the work a little simpler, which is not a bad starting point if you are trying to get your head around how the flow control works for exceptions.

---

Brian Tkatch said:

Personally, i have done this (with a PACKAGE though) by RETURNing a value from the PROCEDURE, and then checking it in the main code:

IF Some_check() = 1 THEN RETURN; END IF;

It's not as pretty, but i found it to workout nicely.

Sure, I'm a big fan of function-oriented design; I think it's been sadly overlooked in the rush to make everything object-oriented.

Restructuring the early tests into functions does take a little work. And it requires some design decisions -- use Boolean values, 0/1, or named constants? how to ensure all cases are handled, maybe use the case statement? OK, I made my function return Boolean values and tested the values inside a case statement; but I can't test the function values via SQL queries, and maybe case will throw a runtime exception if some unexpected value like null comes back. The sample code in Brian's comment doesn't suffer from those problems, but this is the kind of thing I could imagine a junior programmer coding too elaborately.

That's not to say that the resulting code is any better or worse, just that it's now subject to potential new bugs and maintenance (have to document return values etc.) by introducing functions. That's the idea behind the use of exceptions in my previous post, to make the restructured code 100% the same as the original, not 99 44/100ths %.

The Humble PL/SQL Exception (Part 1) - The Disappearing RETURN

Exception handling in PL/SQL is a big subject, with a lot of nuances. Still, you have to start somewhere. Let's take one simple use case for exceptions, and see if it leads to some thoughts about best practices. (Hopefully, this is not the last post in this particular series.)

One common pattern I find in PL/SQL procedures is a series of tests early on...

if not_supposed_to_even_be_here() then
 return;
end if;

if no_data_to_process() then
 return;
end if;

if no_parameters_passed() then
 print_basic_page();
 return;
end if;
...

In real life, these tests tend to use hardcoded constants, queries, etc. that clutter up the procedure and make it hard to follow, rather than descriptive names as in the example above. One simple solution is to move the whole block into its own inner procedure:

check_if_supposed_to_be_here();
check_theres_data_to_process();
check_parameters_were_passed();

These procedures, declared with the procedure ... is ... syntax immediately before the begin of the main procedure, can access all the variables from the main procedure, so they typically don't require parameters. In most cases, you can just lift a block of confusing code from the main procedure, and turn it into an inner procedure with a descriptive name.

However, the return statement complicates things. When transplanted into an inner procedure, it loses its mojo. Instead of cutting short the entire procedure, it becomes essentially a no-op, because now it's at the end of a short inner procedure that was about to return anyway, back to the middle of the main procedure. The solution is to use an exception, which requires structuring the whole business like so:

create or replace procedure big_procedure as
  num_rows number;
  exception skip_normal_processing;
  procedure check_data_to_process is
    select count(*) into num_rows from data_table;
    if num_rows = 0 then
      raise skip_normal_processing;
    end if;
  end;
begin
  check_data_to_process();
  ...do all the normal stuff if there really is data to process...
exception
  when skip_normal_processing then null;
end;
/

Now if you detect some condition that means the procedure should bail out, it really will. If you can anticipate that your procedures might get lengthy enough to benefit from using inner procedures this way, you can plan ahead by using exceptions right from the start, instead of starting with return statements and then turning them into exceptions when you restructure the original straight-line procedure.

One thing that still bothers me is the way the control flow jumps around. The calls to the inner procedures jump backwards, and if any "stop! now!" conditions are triggered, control jumps forward all the way to the end of the main procedure. When I visualize such a structure, it reminds me just a little of spaghetti code. I know that on paper, all is as it should be -- all the reusable / modular code is separated out at the front, all the error handling and termination code is separated out at the end. I just would like to see more real-life cases where such structure saves on maintenance and debugging time, before passing final judgment.

When Backwards Compatibility Goes Too Far

I couldn't help but notice this new article, about holdovers from the earliest days of DOS and even CP/M still showing up in Windows-based development:

Zombie Operating Systems and ASP.NET MVC

Personally, I really enjoyed working on the IBM C/C++ compiler back in the day, targeting Windows 95. They licensed the Borland resource editor and I adapted the RTF-format online help, with no RTF specs, just trial and error. For me, that was the apex of Windows-based technology. Once the web came along, everyone forgot how to write usable desktop apps.

Deconstructing the iPod Shuffle UI

The new buttonless iPod Shuffle, which moves all the controls onto the headphone cord, is taken to task in this article:

The new iPod shuffle: Button, button, who's got the button?

Now, I'm a recent purchaser of the previous Shuffle model, and intuitively I prefer the Play/Pause/Forward/Back/Up/Down controls of that previous model. But I like to take contrarian positions sometimes too, so let me see if I can defend the new Shuffle from a UI point of view.

One thing I notice with the older square Shuffle, is that each time I clip it on requires a mental orientation session. With this jacket, it's clipped on this side; with that shirt, it's clipped on the other side with the controls upside down and reversed; with a t-shirt, it's probably clipped on the bottom, and so the controls are 90 degrees from either of the previous orientations.

This takes a few seconds, which isn't a long time, but it's annoying if the goal is for the player to vanish from my consciousness. I quickly straighten out which button skips to the next track, but if I'm too distracted to visualize how the Volume Up/Down buttons are situated, I might find which is which by trial and error while driving. Plus I may clip it in a different orientation if I decide it's more important afor this outing to turn it on/off, vs. having easy access to the audio jack.

Now, there are a few ways this could be addressed. Some sort of Braille-like bumps to make the buttons distinctive to the touch. A rotating control area -- probably wouldn't help without the bumps. An accelerometer to make the buttons swap functions when it's clipped up, down, or sideways. But all of those just drive the cost of parts up, and still take some getting used to.

Having the controls on the headphones, regardless of whether you like earbuds or not, does make the interaction identical however you clip on the actual player -- in the car, walking, working out, in your pocket, upside down during weightless space training, and so on. (Actually, could you even plug it in to a car stereo without some kind of extra adapter cord with the new controls?)

I think there's a much broader lesson to draw here about Apple's UI principles. Lately, it's introduced lots of things like the Shuffle, the Apple Remote, the single giant button trackpad. The theme here is "Fitts's Law". Broadly speaking, it means that it's easier to select something UI-wise if it's big and nearby. I like the AskTog description better than the Wikipedia article:

AskTog: First Principles of Interaction Design

You see this principle applied a lot in Apple's software UIs. Think about the magnification of icons in the dock as you mouse over them. The thing you might select is directly under the mouse, and it gets bigger so it's hard to miss. Or the software keyboard on the iPhone, where you tap a letter that's very small, yet the screen displays a magnified version of each letter as you type it, so you can slide your finger to adjust if you didn't hit quite the right spot.

Now, the principle is bleeding over to the hardware side of UIs. IBM had the little mouse nub for Thinkpads. Right under your fingers as you were typing, but small and hard to hit or control. The Apple trackpad that's all button, all the time, is a purer expression of Fitts's law. You could even say it embodies a third related principle, that in addition to making the item easy to select, you arrange the UI so that there's only one logical action for this one giant button to perform. So you're not clicking on different parts of the trackpad to do different things, it's always "select"; other gestures, such as double-clicking or the two-finger swipe, are physically different, but again work regardless of the location.

We see this new principle spreading in the world. For example, the video players you see on web sites where the initial screen is a single big Play button. We saw it a while back in hardware with that third-party gizmo that was a dial with a single button. These days when I use a second monitor for Photoshop, I realize it's part of the same trend; this monitor is really a UI control with a single selection -- its own On/Off button, to select when entering or leaving the mode where I need more screen space for menus and palettes.

But anyway, back to this new Shuffle. We've got the controls that are situated in the same location regardless of how the device is positioned and oriented. We've got a single action that the controls perform, with less-frequent variations that depend not on location but on a different gesture (double- or triple-clicking, press and hold). So, although I still like my el-cheapo headphones and cassette adapter for the car, the new Shuffle UI makes sense if considered part of a pattern.