Wednesday, November 4, 2009

Stored Procedures and Ad Hominem Arguments

The question of "Stored Procedures and Triggers" comes up fairly frequently.

Over the years (since the 90's, when stored procedures were introduced to Oracle) I've learned precisely how awful a mistake this technology is.

I've seen numerous problems that have stored procedures as their root cause. I'll identify just a few. These are not "biases" or "opinions". These are experience.
  1. The "DBA as Bottleneck" problem. In short, the DBA's take projects hostage while the development team waits for stored procedures to be written, corrected, performance tuned or maintained.
  2. The "Data Cartel" problem. The DBA's own parts of the business process. They refuse (or complicate) changes to fundamental business rules for obscure database reasons.
  3. The "Unmaintainability" problem. The stored procedures (and triggers) have reached a level of confusion and complexity that means that it's easier to drop the application and install a new one.
  4. The "Doesn't Break the License" problem. For some reason, the interpreted and source-code nature of stored procedures makes them the first candidate for customization of purchased applications. Worse, the feeling is that doing so doesn't (or won't) impair the support agreements.
When I bring these up, I wind up subject to weird ad hominem attacks.

I've been told (more than once) that I'm not being "balanced" and that stored procedures have "There are pros and cons on both sides". This is bunk. I have plenty of facts. Stored procedures create a mess. I've never seen any good come from stored procedures.

I don't use GOTO's haphazardly. I don't write procedural spaghetti code. No one says that I should be more "balanced."

I don't create random database structures with 1NF, 2NF and 3NF violations in random places. No one says I should be more "balanced".

Indeed, asking me to examine my bias is an ad hominem argument. My fact-based experience with stored procedures is entirely negative.

But when it comes to stored procedures, there's a level of defensiveness that defies my understanding. I assume Oracle, IBM and Microsoft are paying kickbacks to DBA's to support stored procedures and PL/SQL over the more sensible alternatives.

11 comments:

  1. If you don't allow the possibility that those who disagree with you could be anything but crazy or corrupt, then it seems optimistic to expect respectful debate from them.

    In environments where the database will be accessed and changed by ad-hoc programs and by many different applications, there are good reasons to keep logic bound unavoidably to the data instead of trying to duplicate it separately in every potential channel of access. In this case, it's important that there not be a sharp and hostile separation between DBAs and developers. If you've never worked in situations like that, though - if your data is always the tail and the application is the dog, instead of the other way around, or if you've never had good cooperation between DBAs and development - then yes, I suppose you've never been in a situation where stored procedures could have helped.

    Or maybe I'm crazy and corrupt, too.

    It would help a lot if approaches like PL/Python were available outside postgresql (or if more workplaces used postgresql). Wanting to use stored procedures shouldn't restrict a person to using a single rather aged and limited language, but in most databases, it does.

    ReplyDelete
  2. I agree with the article.

    I've also come to this realization that stored procedures break the programmer's ability to "fully grasp" the program. This also means that unless the programmer using his language also knows the SP language, then you now automatically involve two programmers, and neither can "hold the program in their head". So then you end up having to write down a lot of documentation (business rules, etc) that too can be misleading, and requires an analyst and lots of writing, and meetings, and now instead of one guy being able to crank out something good and fast you end up with a team creating something bad and slow.

    Why bad? Because no matter how well people think they can write down what's in their brain, putting things down on paper or even communicating verbally creates friction that does not exist if one person is able to just think about the whole problem.

    @catherine: even PL/python is a subset of the real python languange, and does not have the full expressiveness of python + standard library + third party tools and frameworks.

    The other thing is that people think they need to protect the data from the programmer, that they need an abstraction layer. I say: get better programmers.

    ReplyDelete
  3. There are 2 reasons for stored procedures:

    1. Mentioned by Catherine: You have multiple applications written in different languages/platforms and no shared libraries between those accessing the same data and you want to make sure that some validation always takes place. To avoid duplication of code you write it in a stored procedure. This is to make sure there is 1 source for the rules, otherwise you might end up with various apps doing different things.

    2. The stored procedure is doing massive data moves. For example you move a million records form table A to B and it can't be expressed in one SQL statement. By doing it in a stored procedure you can use variables etc, but avoid all that data going out of the database application over the wire to your application server into your application and back.

    Otherwise you are right: Avoid stored procedures as much as possible.

    The people that know the application better are the developers, there is more of them than DBAs, they know their programming language better than whatever language the DB provides, they have better IDEs and debugging tools. Programming in stored procedures and having to depend on the usually fewer expert DBAs slows everybody down. Also if you ever decide to switch database engines for your application you will have to rewrite all your stored procedures. It is better to try to decouple your application from your particular Database implementation.

    ReplyDelete
  4. sigh...developers sure love to hate on the DBAs.

    ReplyDelete
  5. Don't know why someone would argue with you, clearly you are speaking from personal experiences. They are what they are.

    Triggers can definitely get out of hand and are, imho, the most single most abused "feature" of any database. What seems like a god-send rapidly deteriorates into a tangled mess of "magic bullets". If you see a lot of triggers being used that should throw up some serious red flags.

    ReplyDelete
  6. Stored procedures may just be another facet of vendor lock-in. Using one type binds you forever.

    ReplyDelete
  7. My biggest problem with sps is that the become black holes that all business logic fall into.

    Something might start out as a simple stored procedure but sooner or later complex business logic starts creeping in. Then if more than one app requires the same proc critical mass is achieved.

    Sql is great at expressing set based logic but it sucks at expressing business rules.

    And don't get me started on all those code generators that write stored procs "because their faster". Then end up creating the most horifically slow code.imaginable.

    ReplyDelete
  8. If you work in an area where items #1 and #2 are real, then you have a "management" and "staffing" problem, not a "stored procedure" problem. (This is especially true if your developers aren't the ones writing the stored procedures.)

    ReplyDelete
  9. These are pretty arbitrary objections. You could apply complaints #s 1-3 to virtually any group within a software development organization - Business Analysts, for example, or QA testers. It sounds to me like you've just had bad experiences with "database people" and decided to hold a grudge.

    There are perfectly legitimate reasons to create stored procedures - restricting permissions and attack surface, for example, or enhancing performance by caching query plans for queries with complex logic. That doesn't justify jamming everything into a proc, but as with everything else in software development, stored procs are a tool, and can be used and abused like any other tool. How much do you like those sysadmins who write everything in completely unreadable Perl scripts? Or the BAs who code all the business logic in convoluted VBA before handing it over to a "real" coder to re-write?

    I would recommend examining your biases, and giving the DBAs a break, because if you came to me with that attitude, you're damn right I would hold up your project while I tune your stored procs, because now I have serious doubts about your ability to use them effectively.

    ReplyDelete
  10. @ Ira:

    I too have had bad experience with "dba" people then. It seems to be the norm, too.

    So while in theory stored procedures are a great too, in practice they turn things into a complete mess. That's what he's saying.

    Also, and he approaches this obliquely, he states that insisting on stored procedures implies mistrust of the programmer's skills. That can't be good for morale, and demoralized programmers hardly produce good stuff.

    ReplyDelete
  11. There are times when SP have their place (procedures processing lots of data, procedures pulling together several queries/updates into one DB call, rather than several network round-trips and also for use in securing certain permissions, etc., to only within specific function/procedures).

    However... if your procedures are getting written by DBAs, and not developers (who know PL/SQL (or whatever language) you have a bit problem... you wouldn't allow Websphere administrator to insist on writing all Java code, for example!

    As for customers edition stored procedures - yes, this is a nightmare - however, Oracle (and possibly other RDBMS providers) have mechanisms for encrypting code, mitigating against this problem.

    ReplyDelete