Beefy Boxes and Bandwidth Generously Provided by pair Networks
Problems? Is your data what you think it is?
 
PerlMonks  

Re^5: which perl gui toolkit to use?

by BrowserUk (Patriarch)
on Feb 15, 2005 at 18:21 UTC ( [id://431284]=note: print w/replies, xml ) Need Help??


in reply to Re^4: which perl gui toolkit to use?
in thread which perl gui toolkit to use?

Oh I see. I'm only vaguely aware of what maypole actually is (something to do with web apps?).

I don't have any call for such a beast, but if I did, I probably wouldn't use anything so complex. You just give yourself the other type of maintenance nightmare.

That of becoming dependant upon a complex framework, which makes your life simpler--once you get past the learning curve--provided it is bug-free and you can adapt your requirements to the solution it provides.

The problems arise when you encounter a bug or limitation that is critical to your app but doesn't make it to the top of the "needs fixing list"; or you need something that doesn't quite fit with it's way of working, or the authors preferences.

You can always make local patches to correct those problems you encounter, but if they go against the authors preferences or pull it in a direction the author is not interested in going, you then have to integrate your local changes with those made by the author at each release.

For a simple package that is bad enough, but as the dependancy chain deepens, so the number of authors grows, and the problems of coordinating local changes get ever worse.

I once read that the time required for a project grows as T**n; where n is the number of non-overridable decision makers involved. The same holds true for the number of independantly controlled interfaces.

If they do what you need, your laughing, but if they start to diverge from your requirements, or worse, each other, you have a nightmare.


Examine what is said, not who speaks.
Silence betokens consent.
Love the truth but pardon error.

Replies are listed 'Best First'.
Re^6: which perl gui toolkit to use?
by gaal (Parson) on Feb 15, 2005 at 20:30 UTC

    As a principle, what would you suggest as an alternative? It's all well and good when a project happens to be small enough that it fits in nicely with the aptitude and resources available to one person, but as we know not all projects are like that.

    I recall you said something similar recently about POE, and also not as specifically less recently. I sorta agreed with you there — my point was that reusing good design is as good as reusing good code — but sometimes you don't have time to do all the coding yourself. (This is problem is remarkable in UI; and most certainly in web presentation where there's a lot of menial work.)

    You seem to be hinting at a profound distinction between library (or module) and framework. If I'm smelling right, would you care to share your thoughts on the differences?

      You seem to be hinting at a profound distinction between library (or module) and framework. If I'm smelling right, would you care to share your thoughts on the differences?

      You hit the nail dead on.

      Of course, the distinction gets blurry around the boundary, but in simple terms it comes down to this:

      My code calls library modules.

      Frameworks call my code.

      With the former, I structure my application to fit my requirements and call library routines to perform specific tasks.

      With the latter, the framework provides a generic structure for the application, and it calls my code to fill in the details.

      Both are perfectly legitimate, but historically, I have had much greater problems with the latter than the former.

      With the library approach, if the library goes out of mainenance, takes a divergent course or otherwise fails to meet my requirements, I can substitute another library by providing a shallow interface mapping layer between existing calling code and the new library. At worst, I have to write a replacement from scratch--but my investment in the calling code is not wasted.

      With the framework approach, when it takes a course divergent from my requirements or stops being maintained, or when my requirements change to fall outside of it's facilities, my choices are limited. I can petition for changes to the framework to accomodate my needs. I can fork a local version and make my own changes to it, perhaps seeking to feed my patches back into the mainstream with the risk of them being rejected. Or I can move to a different framework, but the chances that my existing investment of code will be easily modified to work with that new framework are low.

      Frameworks have a very difficult and fine line to walk. That between being very generic which makes it most flexible, but also has a tendancy to mean sticking to the lowest common denominator; and being highly specific, which makes for clean, targeted interfaces, but can also limit the range of applications that they are useful for.

      Open source frameworks carry less risk than the proprietory ones that I have had trouble with on two occasions in the past.

      In one case, a small company, 3 or 4 guys got into an internal wrangle that eventually became letigious, stopped development and prevented the company I was working for even attempting to buy the source, because the litigants couldn't decide who owned it.

      In the second case, a quite large and successful company producing a quite widely used framework were bought out by a much larger software house who produced a less successful competitive product. They took the designers, programmers and the good ideas and used them to improve there own product, but squashed the one we were using. The costs and difficulties involved in moving from the original to the competitive product were the death knell of what had been a promising application.

      So yes, I see a world of difference between libraries and frameworks, and I strongly favour the former to the latter.


      Examine what is said, not who speaks.
      Silence betokens consent.
      Love the truth but pardon error.

Log In?
Username:
Password:

What's my password?
Create A New User
Domain Nodelet?
Node Status?
node history
Node Type: note [id://431284]
help
Chatterbox?
and the web crawler heard nothing...

How do I use this?Last hourOther CB clients
Other Users?
Others imbibing at the Monastery: (6)
As of 2024-03-28 08:23 GMT
Sections?
Information?
Find Nodes?
Leftovers?
    Voting Booth?

    No recent polls found