How to develop without an IDE?

[Martin Gregorie]

That is somewhat true.

"designed by committee" is not always a success.

But I do still not quite understand why Gene bring op marketing when I
suggest he use tools recommended by the developer community.

Probably because he thinks my second group of tools are invariably the
result of sales guys pushing their outfit into cashing in with a copycat
development. I'd say that a lot of them are just that, but some aren't.

For instance, I doubt that there was a sales push behind either of
Microsoft's edlin or Wordpad editors, but clearly somebody thought they
were a good idea and published them despite their rather nasty UI (edlin)
and minimal capabilities (both editors). It would be interesting to know
what their development teams used for their everyday editing needs. After
all both vi and emacs predate DOS and edlin by 5 or 6 years and,
regardless of whether you love or hate their UI (which is no worse than
edlin's one), you have to admit that both are extremely capable editors
and initially ran in similarly small memories too.

BTW, I'm not deliberately kicking Microsoft for once: just using their
editors as a examples of rather poor tools that most people on this list
are likely to have used.

 

[markspace]

Microsoft's edlin or Wordpad editors, but clearly somebody thought they
were a good idea and published them despite their rather nasty UI (edlin)
and minimal capabilities (both editors). It would be interesting to know
what their development teams used for their everyday editing needs. After
all both vi and emacs predate DOS and edlin by 5 or 6 years and,
regardless of whether you love or hate their UI (which is no worse than

Just interjecting here: line editors are a much older and simpler tool
than full screen editors. My recollection is that traditionally when
you are writing an OS from scratch, you build the simplest pieces first
to get started, then more complex ones. edlin and the Unix equivalent
"ed" are simple line editors that don't require a working full screen
library. edlin was probably what MS used in the early days to bootstrap
their development. And if you ever found yourself with out a working
full screen library for whatever reason (say you're working remotely on
a teletype) you still have edlin (or ed) in your command line to
configure a few things in the OS.

 

[BGB]

If it is a group for physics/astronomy I would assume so.

Here nobody cares.

If you consider Java import and C include similar in
alt.chocolatecake.baking, then I don't think anyone
will object.

But we will here as this happens to be a programming
group where a certain level of technical accuracy in
relation to programming languages is expected.

I think Patricia summed it up fairly well.

technically, they are not the same: they do different things and work in
different ways, but with their similarity being mostly in regards to the
role they serve.

the cooking analogy would be of substituting ingredients with other
things: the things being substituted are not the same, but in a food
product they may be similar enough. like, something calls for eggs, but
a person uses something like mayonnaise instead, ... no one may claim
that raw eggs and mayonnaise are the same thing, but both may serve a
similar role in certain food products.

 

[JussiJ]

On 4/23/2012 3:38 PM, Rui Maciel wrote:

If you can show me a text editor that is superior in any significant
regard to my IDE.

I might be biased but I think Zeus is a darn good programming editor.

I'd like to see what those features are, and know which editor
provides them

What defines a killer feature is always very subjective and
hard to define.

Jussi Jumppanen
Author: Zeus Editor

 

[Gene Wirchenko]

[snip]

It could be. If that is all that a developer needs and will not
use the other tools, that is fine.

If it turns out that a developer needs to learn another tool --
whether ant, maven, or anything else -- then yes, he ought to. I just
do not see the point of learning a tool that one is not going to use.
Know about, yes; know, maybe. I have sometimes only used a tool once
and never needed it after that. Several versions later, what do I
really know about it? Not much anymore.

So you really mean B.

And my assumption and argumentation was correct.

No, actually, it is not. It can be acceptable for someone to
know only some lesser-known tools -- it may be all that is needed --
but it is not necessarily so.

If I have the choice of spending learning time on a commonly-used
tool that I will not use (for whatever reason) and a more obscure tool
that I will use, I will choose the latter.

I might already have had a look at the commonly-used tool when
tool choice was being decided. I might look into it a bit more deeply
later, but if I see no use to digging in, I will not do so.
Circumstances change, and if the tool is now something of use, then I
will dig in.

[snip]

They are pretty straight.

The mentioned survey put them at 30.8% and 48.2% - I don't
see any problems calling them tools with 30-50% usage.

No, you mentioned one figure. I naturally assumed that it was
for both since you were discussing both.

Sincerely,

Gene Wirchenko

 

[Gene Wirchenko]

On 5/1/2012 1:23 PM, Gene Wirchenko wrote:
[snip]

I don't think anyone has claimed a need to learn any tool under
the SUN.

Of course not. Everyone and his dog says to use his tool. The
collective of all that yammering is one should use every tool under
the sun.

Not at all.

All too often for me. Granted, this seemed to happen more when I
was at uni.

Suggesting tools that are used by millions of Java devlopers
is a pretty good suggestion.

Not necessarily.

Does the suggested tool fit in with what I am doing? Some people
make their suggestions without ever considering this. Sometimes, one
can get wrapped around a pole following such suggestions.

A suggestion that includes how the suggested tool could make it
easier/faster/better to do something might get my attention. If the
person has looked at my situation and actually considers it some, he
might have come up with something of use. I am much more receptive
then.

Do you consider "Java developer community" and "marketing department"
to be similar concepts?

Yes. Both are advocacy groups with their own interests.
Sometimes, the developers are more interested in self-aggrandisement
than in suggesting something useful. There is a reason we have the
term "fanboy".

Otherwise I can not see the relevancy!


You mean go through all relevant tools?

That will be one very long interview!

You mention "all relevant". Previously, you have been on about
commonly-used tools. There is a difference.

Besides, one hardly need go through a checklist of every tool.
Just ask what the interviewee has used. If he does not mention
something of note, ask about it.

You might get "Ant and Maven? Not much. LastCorp had a build
tool that was used across multiple languages. They were very opposed
to breaking things up. Their tool worked very well so after a while,
I went with the flow. No point in arguing. So I just know the bit
from when I was checking them out. I did a few test builds, thought
both tools looked reasonable, but I was not going to be using either.
My personal programming is mainly in C, Assembler, and T-SQL."

Sincerely,

Gene Wirchenko

 

[Gene Wirchenko]

That is somewhat true.

Ha! All too true, all too many times.

Sybase PowerDesigner is very awkward.

Oh, you can get some very nice results from it, but the handling
of method definitions is very awkward. What were they thinking?

The documentation (for the version I used) was broken up into so
many pieces, and it was not readily apparent where the documentation
for a given point was. Again, what were they thinking?

"designed by committee" is not always a success.

But I do still not quite understand why Gene bring op marketing when
I suggest he use tools recommended by the developer community.

Because I know what the word "marketing" means.

Recommendation by the developer community is a form of marketing.

As a purchaser (whether handing over money, buying in, or both),
I have to consider my needs. The needs of the rest of the world are
not of such a great concern. I refuse to join the stampede.

Sincerely,

Gene Wirchenko

 

[Gene Wirchenko]

On Sat, 05 May 2012 20:26:41 -0400, Arne Vajhøj wrote:
[snip]

That is somewhat true.

"designed by committee" is not always a success.

But I do still not quite understand why Gene bring op marketing when I
suggest he use tools recommended by the developer community.

Probably because he thinks my second group of tools are invariably the
result of sales guys pushing their outfit into cashing in with a copycat
development. I'd say that a lot of them are just that, but some aren't.

Not invariably.

Other people have their interests. These interests might
coincide with mine, and they might not.

I will decide on the basis of *MY* interests.

For instance, I doubt that there was a sales push behind either of
Microsoft's edlin or Wordpad editors, but clearly somebody thought they
were a good idea and published them despite their rather nasty UI (edlin)
and minimal capabilities (both editors). It would be interesting to know

I thought edlin ridiculous.

I liked WordPad. If it had had just a few more features, I would
not have had to use Word much, if at all. (Page breaks and just a bit
more formatting would have done it for me.) That would not be true
today, but I still need much less than what Word has. I am still
using Word 97.

what their development teams used for their everyday editing needs. After
all both vi and emacs predate DOS and edlin by 5 or 6 years and,
regardless of whether you love or hate their UI (which is no worse than
edlin's one), you have to admit that both are extremely capable editors
and initially ran in similarly small memories too.

I think edlin is a bit better, but all of them are in my
better-avoided category.

BTW, I'm not deliberately kicking Microsoft for once: just using their
editors as a examples of rather poor tools that most people on this list
are likely to have used.

That was clear.

Sincerely,

Gene Wirchenko

 

[Lew]

do people make a big fuss over "the sun rises and the sun sets" when
in-fact [sic] it is the Earth that is moving?

In fact, in any frame of reference for which neither the Sun nor the Earth is the origin, both are moving.

If it is a group for physics/astronomy I would assume so.

Here nobody cares.

If you consider Java import and C include similar in
alt.chocolatecake.baking, then I don't think anyone
will object.

But we will here as this happens to be a programming
group where a certain level of technical accuracy in
relation to programming languages is expected.

Really, why be an apologist for imprecise terminology? 'import' is an important (pun intended) notion to Java, and confusion of C and Java concepts isa well-known retardant to mastery of Java. The conflation of superficiallysimilar concepts between the languages consistently causes trouble, so it is bad policy to support such error. Instead, assuming a commitment to helping one's fellow practitioners, one should take a stand to firmly reject such confusion.

Given that logic, someone attempting to support sloppiness in the distinction between 'import' in Java and '#include' in C opens themselves to accusations of disempowerment. Validly so, I aver.

 

[sony.gupta32]

what about an xml editor, try Liquid XML Editor (http://www.liquid-technologies.com/xml-editor.aspx)

 

[Martin Gregorie]

what about an xml editor, try Liquid XML Editor
(http://www.liquid-technologies.com/xml-editor.aspx)

Don't use make - ant is better for Java and one control file (build.xml)
can easily handle several non-nested package structures as well as jarfile
builds and javadocs output.

Everything else can be handled via your favourite text editor (java
source, regression test scripts and test data.

 

[Lew]

Just about any Java tutorial, and also on the Ant website.

Ewww! Makefiles don't go well with Java.

Don't use make - ant is better for Java and one control file (build.xml)
can easily handle several non-nested package structures as well as jarfile
builds and javadocs output.

And with a little more difficulty can handle just about everything else related
to builds and kicking off testing.

Everything else can be handled via your favourite text editor (java
source, regression test scripts and test data.

Well, now, "handled" is a loose term here.

Tools like Jenkins/Hudson, Ant, IDEs and various performance and
functional test scaffolds exist for valid reasons.

The OP is right to seek command-line, or rather, command-script
knowledge. All the Java IDEs recognize Ant build.xml files and can
work well with Ant-based build/deployment protocols. It is
important that one does not introduce dependencies on the IDE
itself in those protocols. Then all sorts of standard tools, and
multiple versions of such, will work with the project.

One more thing about Ant. If you really can't do what you want
in it, yet, you can choose to write new Ant tasks in Java and fold
them into your process.

 

[Roedy Green]

Everything else can be handled via your favourite text editor (java
source, regression test scripts and test data.

I used to use SlickEdit for this. You need to install the JDK
see http://mindprod.com/jgloss/jdk.html
You make macros to do common editing tasks.

Use ant for building. I generate my ant scripts with a stomp cookie
cutter program. This mean I can add a project with just a few lines.
For example, last night I noticed I had been signing only the main jar
for JWS. I changed a few lines, and all my ant scripts were updated.

But IDEs are so much more powerful in terms of navigation, tracing,
reformatting.