Python suitability

[Boudewijn Rempt]

>>       From:
>>            boud at rempt.xs4all.nl (Boudewijn Rempt)
>>
>> If you really need to use a variety of databases, then you probably
>> can't rely on stored procedures/packages to do the server-side work,
>> because there isn't a database where those work in the same way. And
>> even so, support for, say Oracle packages, is skimpy outside Oracles
>> own tools. I'm going to install Oracle next week to try whether they
>> can be accessed from Python, though.
>>
>> You really need some middle man, if you go the various databases
>> route, and designing decent middleware is difficult, in my experience.
>>
> Do you have a more precise idea? I am also interested in using middlemen
> (middleware). I evaluated Fnorb and found it perfect.

I haven't done anything with  Fnorb yet - it's an orb, isn't it?, but
what I mean is that you've got to have some part to your application that
ties together all the information and works with it - whether that is
to calculate discounts, to determine customer profiles or calculate the
dioxin contents of a chicken. When you write that part in a proprietary
database, you lose database independence. Just using an ORB doesn't
solve it - you still need to write the code and determine where it
resides. Making it a separate Python application your interface talks
to using Fnorb (or a webserver) could work very well, I guess.

>>
>> I don't think TK is really suitable -

> I am very sad! Then why there are any effort to support TK by Python?

I don't know ;-). I think Tk has had it, it's ugly, hasn't got a lot
of features and is difficult to use effectively - in my experience,
at least.

> I used Qt (but not PyQt) and it is a fantastic library, not only GUI but
> general purpose C++ object library (or rather framework). I would go for
> it.

It all depends on the application you need - if it's an e-commerce app,
you can't use Qt, I fear. But check the bindings out, by all means,
they are extremely neat!

>>
>> I do find Zope a bit slow - even when I browse a Zope database on a
>> machine on my local network.
>>

> Again I am very sad! I wanted to start a new department in our company
> developing ecommerce applications. I am on the point of selecting a
> development environment. Untill now I would voted for Apache+SSL + Zope
> 2.1, but I am also thinking of using Apache+SSL + PHP or ApacxheSSL +
> mod_perl.
> Do you not think Zope would be a good platform for our projects?

Zope is perfect for e-commerce applications - it's a domain it's well
suited for, especially if you don't try to imitate a local app
interface with it. That's a matter of raising wrong expectations. If a
user sees a picture that resembles the widgets his common apps are
built with, he will expect the same responsiveness. And you won't get
that from a webserver, not even on the local intranet. But it really
depends on the project, on the type of app you're making.

> Do you have more arguments for not taking WxPython enterprise-ready? The
> lack of database-ready widgets does not bother me, as I access databases
> through CORBA (Fnorb) servers.

Well, it's multi-platform, but requires a non-standard version of GTK
to build on Unix - meaning that simple-minded users have to jump
through all kinds of hoops to install your app. And I somehow
don't think Windows users are all that ready to install another
extra library - but if your group of users can support it, then
it's worth trying.

> How many time (days or weeks) do you think is needed for a good C++
> programmer to become a good Python programmer?

Not many weeks, a month or so, I guess - but that's just a guess. I've
never seen more than two good C++ programmers in the same room! So, if
you've got a bunch of them, hold them tight and don't let them go. They
might baulk if they have to use a sissy script language, though ;-).

> We have at our company a department using Microsoft tools. They produce
> loss for five years. I am a bit afraid of using any Microsoft tool.
> However, our department uses no Microsoft development tools at all: we
> produce outstanding revenues for five years. Microsoft tools create revenue
> only for Microsoft!

It's only a tool... Look closely at the development process, and the
quality of the personnel. If your department is staffed with people
who can work with Qt and C++, it's no small wonder they can write code
in time, in budget and with a good quality - most VB shops are staffed
with beginning programmers, and inexperience costs money.

>>
>> I've never seen C++ or Java used for large-scale database applications.
>> Most projects I've seen use Oracle Forms, Visual Basic or Powerbuilder
>> or things like that.

> I was also involved several of this kind of projects: never again! I would
> never use again any proprietary language or development tool.

Oh, horses for courses - if you want that e-commerce application running
within half a year, I'd go for Oracle 8i, and get it over with. Sure,
it's proprietary as hell, but it works. Or Zope, but Oracle has a gentler
learning curve... Though that's not saying much. But there are very,
very nice sites built with Zope. The KDE news site: www.kdeforum.org is
responsive, attractive and a good all-round example, I think.

> This sounds interesting! I am also interested in PyQt.

Try it - it's free ;-). http://www.river-bank.demon.co.uk/software/
is the address for the bindings.

> This is quite serious to be only a hobby project. Several corporations
> would pay a lot of money having a programmer being capable of designing and
> implementing an application using this architecture.
> I guess yoiu have to create VB programs during work hours! Poor man, I
> would not do that! You should change your employer!

I'm not saying I can pull it off! I'm just _trying_ to get it to work.
I know it's the right way to go, though. I've been through enough
client=server and three-tier debacles to have finally found the
Right Way.

>> My bosses don't want
>> another programming tool in the shop, after the debacle with VB 5.0.

> Your bosses may be complete idiots! VB5 and any other VB are rubish!

I can see their point. They've just made a mistake that cost about
five million dollars, a mistake they need some seriously expensive
external consultants to force them to admit they made it, and they
are very, very shy of making another decision. Better say it's a
no go. What doesn't happen can't go wrong...


> A company using VB would not be a competitor for us. I am sure we are not
> the only company focusing on leading edge technology.

My company isn't exactly innovative... Nor competritive. After
the recent debacle the mantra is 'we're not a software company,
we don't have to strive for excellence in systems development,
it's not our core business'. 

-- 

Boudewijn Rempt  | http://denden.conlang.org