java or C++ ??

[patrick]

Yes I posted on comp.lang.C++ 1/27/2007.

patrick

 

[Chris Uppal]

I checked at the time of my first reply that this
was not x-posted,or multi-posted, but apparently
the C++ post was 'late in showing up' in the
GG listing.

Perhaps a question of this nature would have
been better *x-posted* to the *advocacy* groups
of Java and C++.

I think not. At least not unless you think the question is only fit for a
advocacy grouo (I don't, myself).

This would be a reasonable example of a case where multi-posting is to be
preferred over cross-posting -- because the overlap of the C++ and Java
fanatics would almost certainly generate so much heat as to drown out the more
interesting exchanges between the moderates in both camps.

Slogan: /anything/ is better than a flame-war -- even multi-posting.


BTW: one thing that nobody has mentioned here is that (assuming the C++ version
was written and worked), there are delivery challenges for "native" programs
too. Toolkits such as GTK+ and wxWidgets involve failrly large downloads (if
you can find a binary download at all), and may have other dependencies which
you wouldn't want an end-user to have to sort out. Of course, you could write
to the OSs native libraries (presumably pre-installed for the most part), but
that approach is not without its little problems too...

-- chris

 

[Kai Schwebke]

why are you asking programmers a marketing question?

Choice of technology should be decided by management,
not from marketing.


SCNR
Kai

 

[Randolf Richardson]

No. In it is very likely to be done in java considering all the comments
and that nobody here favours C++ and that I already know java.But I am
not sure that is a good idea still.

I was being a devils advocate for C++.I wanted to see what the issues are
and how other people see them. But I would be worried that people in love
with java might be biased in their views.

I asked the same question on a C++ group.

Did they counter my points about third-party libraries, such as OpenSSL?

This most intersting comment there suggested I develop in C++ after a
java prototype:.
This is the quote:
"But a large chunk of the dial-up user
base just wont wait for the download to be over. Thats not something
you should dismiss so easily.

Did they explain why you shouldn't dismiss it so easily? If so, what was
the reason they provided?

Also, does your target audience rely heavily on dial-up internet
connections? If your audience is in North America, then you don't have
very much to worry about since high-speed internet connections are so
common-place that many users (based on my experience) don't even seem to
know that non-high-speed options like dial-up even exist.

Anyway, it seems you don't really have much of a choice. You only know
Java so just go ahead and at least write prototype in that language.
Hire another developer when the project takes off."

You're making an assumption about people here, and that is not an
objective way to look at things.

 

[Andrew Thompson]

This would be a reasonable example of a case where multi-posting is to be
preferred over cross-posting -- because the overlap of the C++ and Java
fanatics would almost certainly generate so much heat as to drown out the more
interesting exchanges between the moderates in both camps.

Slogan: /anything/ is better than a flame-war -- even multi-posting.

I disagree with your conclusions (an occasional
flame-war can be fun to watch, and I think it
helps make clear* any 'seething hatreds' between
group members - hatreds that affect their posts
and replies on other threads, but they are usually
too 'polite' to mention.)

Plus I detest multi-posting.

OTOH - thanks for your alternate perspective,
it did somewhat surprise me that (any) people
might prefer multi-posting to a (potential)
flame-war.

Andrew T.

 

[Luc The Perverse]

I disagree with your conclusions (an occasional
flame-war can be fun to watch, and I think it
helps make clear* any 'seething hatreds' between
group members - hatreds that affect their posts
and replies on other threads, but they are usually
too 'polite' to mention.)

Plus I detest multi-posting.

OTOH - thanks for your alternate perspective,
it did somewhat surprise me that (any) people
might prefer multi-posting to a (potential)
flame-war.

I regretted crossposting from this group to sci.crypt when a flame war
erupted between Tom St Dennis and Roedy Green

 

[Andrew Thompson]

I regretted crossposting from this group to sci.crypt when a flame war
erupted between Tom St Dennis and Roedy Green

OK. Chris' advice, together with your own
bad experience, suggests it might pay for
me to reconsider my attitude to x-posting/
multi-posting in the circumstance that it
crosses 'major hierarchies'* or could lead
to a potential flame war**.

* The folks on the HTML hate people x-posting
(for example) an JSP/HTML problem to c.l.j.p.
& comp.infosystems.www.authoring.html(?)
stating that x-posts should not be made across
'major hierarchies' of groups - I am not sure
quite what part of the group name represents
the 'major' hierarchy, though I usually consider
the comp.lang.java.* groups as a single
'major hierarchy'.

** I recongise that while flame wars can be fun
(for me at least) to watch, they can also become
very boring, and wasteful of bandwidth***, if the
participants degenerate into meaningless attacks
on each other, rather than focusing on the technical
points.

*** I also vaguely wonder if flame-wars turn out
to be the 'last straw' for technically proficient
people who get frustrated with the group, or
usenet in general, and 'opt out'.

Andrew T.

 

[Luc The Perverse]

**Massive Snip**

*** I also vaguely wonder if flame-wars turn out
to be the 'last straw' for technically proficient
people who get frustrated with the group, or
usenet in general, and 'opt out'.

I think you are taking this all way too seriously - Although you did make a
very valid point. No multiposting between related groups (like two java
groups). People who are going to get fed up with UseNet are going to get
fed up and leave generally without regards to the actions of any particular
individual.

Usually I will have one issue involving two groups - and I will make
slightly different versions to ask each group - then it is not technically
an identical post.

And yes, this is much more work than typing up a single message and sending
it all over creation.

In this particular issue neither C++ fanboys nor those of JAVA would concede
that there might be a valid point on the other side of the fence. From the
point of a user it can be hard to differentiate fanboy rants from legitimate
advice. (I still think Java was the right answer in this particular issue.)

The real problem comes in not know what question to ask, and the more open
ended the question - the more likely it is to be ignored . . or cause
irritation among regulars.

Usenet patrons generally enjoy advanced eclectic discussions - as those who
don't understand them can be left out. Exceptionally trivial, beginners
questions will get answered by someone, even if they too are generally a
novice. It's the people in the middle that suffer - I recall several times
asking questions, where the answer should have been obvious to a moderately
skilled programmer - but I wasn't asking it in the right way (wrong
semantics) and so my question was dismissed. It's no one's fault - and
nothing has changed. The best way to learn about something, anything - is
many hours of book work, private tutors, paid instructors who have incentive
to try to decipher the gibberish of the uneducated pupil.

I must be tired . . . I don't even know what I'm blabbering about.

 

[patrick]

Did they counter my points about third-party libraries, such as OpenSSL?

That was not posted there. My only comment on that is that openSSL is only
a 3MB download I think so still small.

Did they explain why you shouldn't dismiss it so easily? If so, what was
the reason they provided?

No explanation given. presumably because getting downloads to dial up land
is still important outside the US.

Also, does your target audience rely heavily on dial-up internet
connections? If your audience is in North America, then you don't have
very much to worry about since high-speed internet connections are so
common-place that many users (based on my experience) don't even seem to
know that non-high-speed options like dial-up even exist.

That is interesting. Certainly they would be a big target and if they all
have broadband access I would tilt towards java.

OTOH 16MB is still a quite slow download on a 1MB broadband connection
(common in Europe)
and even on 3MB which I have myself it is a 7 minute download. that is still
a big deterrent when there
is competition with 1 or 2 minute downloads.

You're making an assumption about people here, and that is not an
>objective way to look at things.

I agree with you. that was somebody else's comment, not mine.

patrick

 

[Randolf Richardson]

That was not posted there. My only comment on that is that openSSL is
only a 3MB download I think so still small.

Consider also that OpenSSL, even if bundled in your package, will still
need to be installed somehow. If you automate the installation (which
will be extra work for you), then you'll need to make sure you don't cause
problems for someone using a different version that might already be
installed on the system.

SSL is just one example though. You'll need to make considerations for
all other dependencies as well, and the requirements may increase with
each platform (Operating System) you need to support. With Java, most, if
not all, of what's needed tends to already be included anyway.

No explanation given. presumably because getting downloads to dial up
land is still important outside the US.

You should ask them why they think dial-up is so important.

Consider also the amount of data downloaded via many web sites these days
(image files are getting larger not only due to higher resolution and more
colours, but also because less emphasis is being put on compression the
image data due to the popularity of high speed internet connections among
both end users and businesses, and in many countries).

That is interesting. Certainly they would be a big target and if they all
have broadband access I would tilt towards java.

Don't misunderstand me -- there are still people using dial-up, it's just
that there aren't typically enough to warrant worrying too much about.
Add to that the fact that dial-up users are so accustomed to waiting for
downloads of web pages, eMail messages (which often include both
plain/text and html/text versions, plus who-knows-what-else), and updates
for various software, that a 16 MB download is a fact of life that they
just cope with (I know a lawyer who is still on dial-up everywhere but
home, and he just patiently puts up with it when he's away).

OTOH 16MB is still a quite slow download on a 1MB broadband connection
(common in Europe) and even on 3MB which I have myself it is a 7 minute
download. that is still a big deterrent when there is competition with
1 or 2 minute downloads.

Something's wrong with your broadband connection then because a 3 MB file
shouldn't require 7 minutes to download at 1 Mbps. I'm connected at 10
Mbps at home, and a 2.5 MB download from a decent server just completed
for me a few minutes ago in less than 8 seconds. If my speed was 1 Mbps,
then that download should, in theory, complete in less than 80 seconds.

If I was that worried about download times, then I'd very seriously
consider writing my code in Assembler and use very good de-compression
libraries for graphics and data (which I'd compress before-hand in the
most optimal way).

I agree with you. that was somebody else's comment, not mine.

When quoting others, it's important to indicate that clearly, otherwise
people will assume the comments are your own.

I'm glad that you agree, and hope you'll make the right choice for Java
or C++ for what you're going to do. Writing your prototype in Java first,
and then moving to C++ later, as was mentioned in a different message in
this thread (you may have brought it up, but I don't remember off-hand),
will be problematic in a very time-consuming way because these two
languages have many fundamental differences. It's perfectly okay to use
the same language for both prototyping and everything else, which will
definitely make life simpler for you.

 

[Randolf Richardson]

On Tue, 30 Jan 2007 00:48:53 -0800, Luc The Perverse

The real problem comes in not know what question to ask, and the more
open ended the question - the more likely it is to be ignored . . or
cause irritation among regulars.

Usenet patrons generally enjoy advanced eclectic discussions - as those
who don't understand them can be left out. Exceptionally trivial,
beginners questions will get answered by someone, even if they too are
generally a novice. It's the people in the middle that suffer - I
recall several times asking questions, where the answer should have
been obvious to a moderately skilled programmer - but I wasn't asking
it in the right way (wrong semantics) and so my question was dismissed.
It's no one's fault - and nothing has changed. The best way to learn
about something, anything - is many hours of book work, private tutors,
paid instructors who have incentive to try to decipher the gibberish of
the uneducated pupil.

I must be tired . . . I don't even know what I'm blabbering about.

Please continue, it was very interesting!