Hungarian Notation

It has been a while since it was last discussed online, but it has this nasty habbit of returning over and over again: Hungarian Notation.

And I'll admit, my presentation on NavTechDays contains a slide about the subject too.

There are as many opinions as there are developers.

So what is it about, this "Hungarian Notation". Basicaly it's about variable naming. There is even a page on MSDN about the name.

It's called Hungarian Notation because its inventer was a eastern (or central) european guy so his collegues started to call it Hungarian Notation. There, you know it.

The principle is about the value of adding type identifyers to your variables. Something that has been discussed at any Nav Partner probably over and over again.

Examples:

  • boolBusy  : boolean
  • txtInitial  : text
  • cApples   : count of items
  • flagBusy  : boolean (flag)
  • iSize   : integer
  • decPrice  : decimal
  • pFoo   : pointer
  • arrStudents  : array
  • fnFunction  : function name
     

Problem is, in order for this to work, there has to be a standard that everyone in an ecosystem uses.

Whether you like Hungarian Notation or you don't, fact is that Microsoft Dynamics NAV does not have that, so every attempt to introduce it will undoubtly fail, unless you rewrite the base application and never interact with other NAV applications in the ecosystem.

Other problem is, many partners and/or developers do use it, but all with their own way, so the problem gets even worse since code get's moved around and people switch jobs or partenrs swicht systems. AARGL. Stick out tongue

Microsoft has stopped promoting this methodology years ago and Navision has never used it.

Variables used in Dynamics NAV have a tendancy to be self explaining, like Amount, which is most likely to be a decimal right? And CustLedEnt must be a variable of the record Customer Ledger Entry.

When I started programming the company I used to work for used a Data Dictionary which only allowed you to use variable and field names that where in this Dictonary. When you wanted to use something else you needed to request this name at with the DBA guys. Needles to say this was a COBOL system even though I'm only 34 years old. Cool. Not everything in Cobol was bad you know, except waiting 15 minutes to compile if you're lucky. BTW: This company did not use Hungarian Notation either.

Only exception IMHO are extra explaining identifiers that cannot be derived from the self explainingness of the variable, such as Temp or Buffer for a temporary record variable or p for a pointer in a function.

Anyway, I'll pobably cause some reactions since this is, like I said, somthing that has many, many opinions.

So what are my reasons not to believe in Hungarian Notation

  1. We are developing on top of a huge base application that we all share and does not use this notation
  2. We are part of a wide ecosystem of partners that cross-sell each others solutions that get combined with this base application. Our customers expect us to do this in a seemless way that allows ease of maintenance and upgrade
  3. I do believe in the usefullness of Hungarian Notation in some languages but for some reason it does not improve the readability in Dynamcis NAV

Cheers.

This is published in the contents of my presentation at NavTechDays and with regards to the Partner Ready Software project.

Comment List
Parents
No Data
Comment
  • Disagree. There is an advantage in using a different naming standard, than what Microsoft is using in NAV. Otherwise you risk an object merge (like when upgrading to a new version, or adding an add-on) causes strange problems. This happens if you have created a new local variable, but the upgrade adds a new global variable with the same name or vice versa. This can cause unpredictable behaviour, that in worst case isn't noticed during test.

    I therefore recommend using a different naming standard, and preferbly a standard that actually adds value when working with the code. I agree that you very rarely are in doubt about the variable type, so Hungarian isn't my favorit choice. I prefer a standard indicating if the variable is global, local, parameter or a return-value. This is not shown in the status bar in the C/AL editor in NAV, unlike the variable type. My current company is therefore prefixing all our variables with g_ l_ p_ or r_. Works like a charm, and has even caused developers to be better at making local variables :-)

Children
No Data
Related
Recommended