POST J*** Notes
OpenLook User at 34th IETF (ietfol@mci.ietf.org)
Sun, 03 Dec 1995 13:13:16 -0600
This is a multi-part message in MIME format.
-----------------------------2448181694865
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
http://www.io.org/~mentor1/phpl.cgi?jn951124.html
-----------------------------2448181694865
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline; filename="phpl.cgi"
<BASE HREF="http://www.io.org/~mentor1/phpl.cgi?jn951124.html">
<!DOCTYPE HTML PUBLIC "-//Netscape Comm. Corp.//DTD HTML//EN">
J*** Notes
Mentor Software Solutions presents:
| J*** Notes
for November 24, 1995
J*** Notes is a weekly summary of
information about the Java language and the HotJava(tm) browser, culled
from the mailing lists and news groups. J*** Notes is
in no way connected to Sun Microsystems, originators of Java and HotJava.
Due to time constraints, not all that transpires in the mailing lists and
news groups is reported. What you see here reflects my own eclectic
interests. While we attempt to give a fair and unbiased portrayal of the
exchanges, we offer no guarantees that the information contained here is
accurate, useful for anything other than your entertainment, or that it
won't cause your hair to fall out. We invite helpful comments.
Java, HotJava, and Sun are trademarks of Sun Microsystems.
All other product names mentioned herein are the trademarks
of their respective owners.
|
J*** Notes is copyright © 1995, David Forster.
Individual items are copyright © 1995, David Forster,
Wm D. Clendening,
Mike Shaver,
Rehan Zaidi, and
R.E. Fikki.
Sections are separated by "coffee bean bars."
To return to the table of contents, click on any coffee bean bar
(except this one).
The Contributors
J*** Notes is produced by David Forster (de facto editor), William D.
Clendening, Mike Shaver, Rehan Zaidi, and Rocky Fikki.
We're still trying to get the kinks out of what we're doing, so please bear with
us for the next few weeks. (It ain't easy slogging through about 800 messages weekly,
and then organising what's summarised later.)
Wm D. Clendening
(wdclen@pobox.com) is finishing his PhD
in chemical physics. He became interested in Java because of a long standing
interest in personal agents. Initially, he was bewitched by the hype of
Telescript; but has recently seen the one true light - Java. He hopes to make
Java an integral part of how he earns his bread.
Mike Shaver
(shaver@neon.ingenia.com)
is a network applications and security
specialist for Ingenia Communications
Corporation in Ottawa, Ontario, Canada. He's had a soft spot for
Java ever since he first encountered it in April of 1995, and has been
trying to work it into every nook and cranny of his life since. A
recent high-school graduate, he's going to be attending Carleton University in January, in a
Program To Be Named Later. His other projects including Linux kernel
hacking (especially the networking), Ultimate and theatre.
Rehan Zaidi
(arzaidi@novice.uwaterloo.ca)
is an Electrical Engineering student at the
University of Waterloo
in Waterloo, Ontario, Canada. He is presently employed as a co-op student by
Bell-Northern Research in Ottawa. Aside from his academic interests in
telecommunications and semiconductor devices, Rehan spends too much time
trying to keep up-to-date with Internet applications like Java.
R.E. Fikki
(rfikki@paradise.littleblue.com)
recently started his own (Web Presence/Design Firm: WEB
CREATIONS) over the past few years he has spent much of his time
dealing with internet related applications. Having focussed his attention on Java,
he is now looking forward to the future capabilities that it will bring to the WWW.
He has a Bachelors Degree in Chemistry, and worked as an environmental consultant before
fine-tuning his internet related skills.
And of course, I am David Forster, the bloke what started this thing.
I've been studying and working as a computer professional for almost 20
years. The last five years I taught at the
Department of Computer Science
at York University, but now I consult for a living as the proprietor of
Mentor Software Solutions, and Java
is fast becoming my raison d'etre in that regard.
So, if anyone needs any work done, or has some good leads...
let me know! (phone: +1-905-832-4837).
... and a Different Request
Many people have volunteered to help with the Notes, and we're currently going
through the list of people who have offered. We will probably eventually need
more help still. We'll put out another request when that time comes. We have
another need right now, and that is to find out how we can serve you better.
Our survey has been delayed one week, pending our making up our minds what
to ask you... Please tune in again next week when we'll have a reader survey
prepared for you all,
which we'd very much like you to fill out then.
And as usual, please let us know
what you think about it.
- Announcements
- Features
- Discussions
- Bugs and Warnings
- Comments
- Queries
- Class Exchange
- FAQ candidates
- Help wanted
- Errata
- vanzyl@silas.cc.monash.edu.au
announced an Australian JUG, and a Java Programming Contest ("Win a Sparc 5!").
Details at http://sunsite.anu.edu.au/ajug
(for best results, use a Java-aware browser, especially if you want details on
the programming contest!).
- elharo@sunsite.unc.edu
is looking to add a list of
JUGs to his FAQ,
and asks people to let him know of any groups he hasn't got listed
(or if you want to be the contact person for a new group in your area).
Please give him the location and contact information, at a minimum.
- A JAVA-MAC mailing list has been formed in order to discuss software
development using Sun Microsystem's Java language on the Apple
Macintosh computer (topics: tools, techniques, solutions, tech notes, etc.).
Send "SUBSCRIBE JAVA-MAC your@e-mail.address" in the body of a mail message
to majordomo@natural.com to subscribe.
- 8patches@cftnet.com (Matt Perez)
is looking for someone who can comment
in Java and the changes it will spur in layman's terms. "Your comments would likely
show up in national print media." says he.
- john.zukowski@rssi.com announced the first meeting of the Mid-Atlantic JUG in
Vienna, VA, USA, with future meetings to rotate around the area.
See http://www.rssi.com/info/majug.html
for more information.
- mangalam@uci.edu has a
WAIS archive
of various newsgroups, including comp.lang.[perl, java, tcl], that can be
accessed without the need of a WAIS proxy (thanks to Johnny Goldman's nifty
fwais perl script).
- The Research Triangle Park, NC Java Users Group has formed. For more
information contact Rayme Jernigan
The real-time discussion continued, with lots of Ada-related mud-slinging.
The major point should not be whether or not Ada-XX is good/great/whatever,
but whether Java is suited to real-time work, how it can be improved, and
what you'd want RT Java for. kelvin@iastate.edu provided some light on this:
He doesn't think that Java currently provides the necessary support and has
proposed some RT Java facilities which he summarises as:
- Support for scheduling of periodic, sporadic, dynamic, and ongoing real-time tasks....
- Support for static analysis of worst-case execution time ....
- Support for measurement of average-case execution times....
- Support for real-time garbage collection.
- An integrated real-time executive....
RT applets would be able to negotiate with the real-time executive for enough
resources, and developers could write them to support variable service quality
according to the result of the negotiations.
As for what RT Java would be useful for, he suggests
pen computers, automotive
diagnostic computers, video games, telephone switches, personal digital
assistants, air traffic control, in-vehicle navigation systems,
interactive tv, voice I/O subsystems, tv-top multimedia controllers,
digital teleconferencing software, military defense surveillance,
robotics, and many more applications. The sophistication and market
penetration of embedded computing are rapidly expanding. And there
are already many more embedded computers in the world than there are
"personal computers."
Encryption and Authentication
There was a lengthy exchange on encryption and authentication.
greg@qualcomm.com has implemented a few of the RSAREF functions, although the lack
of unsigned arithmetic was proving a hindrance. He has put the code online.
Some of the discussion concerned the design of the classes, but it was
rather cryptic :-).
david.hopwood@lmh.ox.ac.uk mentioned Eric Young's SSLeay libraries
(http://www.psy.uq.oz.au/~ftp/Crypto/)
which are free of export/key length problems (they were
written in Australia), and are free for commercial and non-commercial use.
(This, apparently does not alleviate the anguish induced by the US government
upon its citizens because of hassles with patent licenses and the categorisation
of encryption software as "munition"! Mik Clarke (an Aussie by his address)
claimed that the Australian government also has export restrictions.)
The cost of training
There was a heated debate regarding the fees sun is charging for its
training courses. The debate was centered primarily around the fact that it may
or may not have been appropriate for them to charge such high fees, if any
at all while the language was not completely formulated yet (beta). Comparisons
between Microsoft and Sun showed a similar approach to such issues (no charge for
promotional sessions, charges for courses).
- capo@ipvvis.unipv.it wonders if the "strange behavior" he sees when clicking
on a one element Choice list is a bug.
- kaj@medialab.ericsson.se says there's a bug in GridLayout that causes a
divide by zero for the case of a panel with no elements.
- mlt.software@ortel.org found that both the appletviewer and jdb seem to want
to connect to the net before continuing. One person speculated that it was
trying to get the name of the host from the DNS; another thought it was trying
to find out if the source to be accessed had been modified since the local
copy had been brought in last.
- Java-based JPEG support in Netscape may not work properly. dwight_aspinwall@theriver.com
reports problems with N2B2, but not with the appletviewer.
- ser@jersey.uoregon.edu reports a documentation bug: The latest Java spec. (p. 21)
gives the range of a byte as -256 to 255; should be -128 to 127, inclusive.
- kenno@servtech.com finds that the colours on a Netscape page with three
Netscape frames on it (each with a different colour) change when an Applet is
displayed in one of them.
- steve_lewallen@taligent.com found a few bugs/misfeatures in WinNT 3.5/beta:
no transparent components (alpha had 'em), mouse_enter/exit problems: no mouse_enter
is generated when leaving a child component encircled by its parent and other
weirdness, null pointer exceptions with lots of components in a GUI applet, and
Javadoc does not create package documentation.
- pat@icon-stl.net notes that the beta docs say the ALIGN attribute of
an applet is required and wonders if this might be a mistake.
- AvH replied
that this is a mistake and he'll make sure that it is corrected.
- mattias@cdt.luth.se reports that the method
isAbsolute() thinks a
(UNIX) link is an ordinary file, so getAbsolutePath() doesn't resolve
links. To avoid a security nightmare, he wants to know how to detect/resolve links
in the beta release.
david.hopwood@lmh.ox.ac.uk suggests an untried kludge: read the file's permissions,
and then dereference it yourself.
- There's a rumour that the Java "jit" compiler will be available in the first
quarter of '96.
- The
GridBagConstraints and GridBagLayout classes
are not missing from the JDK beta distribution. They are in java/src/java/awt
and just need to be compiled.
- parrt@lonewolf.parr-research.com is designing a Java code generator for
PCCTS.
- A query from last week asked how to rig compiled classes to be executed as
commands without using "java" as part of the command. mellis@tiac.net found
that a Win95 solution of associating the .class filename extension falls prey
to some mindlessness on the part of Windows, and worked out this
incredible hack, um, er, workaround. David Hopwood
pitched in with a suggestion to just write a shell file for UNIX.
- jdart@netcom.com wanted to clarify that it is indeed possible to use native code
from within an Applet *as long as the class referring to the native code and the
native code itself are located on the client's machine already*. He was concerned
that recent messages may have led people to believe this was only possible in an
application. Your CLASSPATH and PATH and/or LD_LIBRARY_PATH do need to be set to
point to this code, of course. He warns of an apparent bug in Windows Netscape
(maybe others?) which causes "an unnamed exception at the point where
System.loadLibrary is called" if you create threads in your applet.
mrm@puffin.Eng.Sun.COM follows this up, noting that Sun considers this an okay
way to load third party DLL's. She does urge us to put only trusted .class files
in these directories, and to keep the paths used as simple as possible.
Her note nicely sums up some of this.
- simon@lia.di.epfl.ch has the javac compiler (but not the appletviewer or jdb)
for SGI IRIX 5 working, see
http://liawww.epfl.ch/~simon/java/irix-jdk.html.
- gary@kampai.euronet.nl wanted to see an extension to the language to include
some low-level bit-twiddling functions like "rotate", which he claims would be
useful in encryption, graphics, etc.
- And on the lighter side, david.hopwood@lmh.ox.ac.uk recommends
http://www.digicrime.com/
for the security-impaired. Good substitute for a
late-night read when you've run out of horror stories.
- It seems more people are now interfacing with Oracle, or planning to. References
include http://weblogic.com/,
http://www.algonet.se/~etoile/p1/index.html
(a company called Vincent Engineering), to say nothing of three products just released
by Oracle, including a Web Server, Listener, and Agent.
- carrato@servtech.com comments that running Java animations on a 486SX/33
with 20 megs of memory and Windows NT 3.51 is slow -- animations flicker
severely. It's better but still jumpy on a Pentium 60. [Second data point:
animations are fairly smooth on my 486 DX4/100 with 16 meg, but what is
"smooth?" what "jumpy?"] (He goes on to ask about performance issues, but that
horse is dead.)
- Axel Boldt called for people to drop Java and move to Gnu's GROW project;
most of this thread was irrelevant to us (dealing with suggestions sometimes
verging on the paranoid that we were Sun's dupes, etc.), but one comment from
bhogan@rahul.net was relevant, urging the Gnu people to start developing a
"gava" from scratch, to avoid possible future legal hassles.
- robert_dant@msn.com reports that Borland is just getting started on their
Java development tools and a beta will not be available for awhile.
- daniel@elseware.com says that the Container class needs to
have additional methods which can be used to replace specified components
with other components
- reid@cps.msu.edu wants to know from folk at
other universities if they are using or plan to use Java.
- Various queries related to layout managers appeared: do they work as documented?
Symptoms included components not appearing in their proper places or not appearing
at all. [See last week's issue about a question on items dynamically added or removed
from a layout.]
Another question was: do you need a layout manager in place for text positioning on
labels to work?
- bhickman@pierian.com wants to load a gif file in an application (*not* an
applet, so the stub stuff is missing).
flar says it was an unfortunate oversight
which may not be fixed for a while. A workaround is to use "one of the undocumented
and Sun-specific classes sun.awt.image.FileImageSource which takes a
String filename in its constructor, or the sun.awt.image.URLImageSource
which takes either a String href or a URL object in its constructors.
You would need to do one of: Component.createImage(new FileImageSource("/tmp/myimage.gif"));
Component.createImage(new URLImageSource("http://myhost/a.gif"));
Component.createImage(new URLImageSource(url));"
- bcr1@inel.gov needs help getting a UDP Socket going.
Can the
sun.net.DatagramSocket be used? If so, where is the
documentation?
michael@earthlight.com suggests that classes in sun.net.* (like that mentioned)
will do UDP sockets, but would also like to have some documentation.
- bharat@Corp.Sun.COM needs to know how to trap events
BEFORE they get to a TextArea object?
dufourd@elec.enst.fr replies with
public void postEvent(Event e) {
super.postEvent(e);
if (e.id == Event.MOUSE_ENTER) requestFocus();
}
If your component does this, it gets the focus when the mouse is on it.
- rbk@ibeam.jf.intel.com wants more flexibility in dealing with components, noting
that
java.awt.Button only provides an "action" hook (and no mouse
down/up events), and java.awt.TextField and
java.awt.TextArea don't provide
access to keystrokes. Are there plans to open this up more in a later release?
- daniel@elseware.com wants the contents of one choice list (not the selected
item) to depend on a selection in another choice, and asks if there's a neater
way than removing the old choice component and replacing it with a new one.
- michel@mmania.com was surprised when the code
b = new ( "Class"+"A" );
produced the error message: test.java:7: 'new(...)' not supported.
This is obsolete (he was working from the Alpha spec.). Something like:
b = Class.forName("Class"+"A").newInstance(); will work.
- kristin@kristin.com asked last week if different parts of applets can be loaded
from different URLs
This week simonr@neepneep.uk.su>
Transfer interrupted!
-----------------------------2448181694865--