I took a look at AspectJ awhile ago and instantly spotted the hidden horror show for implementation: There is no test you can write to assure your aspect was applied appropriately. (at least there wasn't when I looked).
Apparently nobody sees this as an issue?
The application of aspects is essentially a blind process before compilation occurs. The best you can do is place errors in your aspect and break the compile.
I can't imagine turning this tool loose on a development organization.
Seems to me this whole issue is a direct result of MS's tarnished brand. Why bother doing research to find out if this weeks security hole is bogus or not? Microsoft's brand is so coupled with "security compromise" you don't need to prove the case anymore to attain public credibility.
I had a similar problem with a guy at that was placing the mouse on the monitor and following the pointer around with the mouse on the screen.
Took about 2 hours on the phone to figure that one out.
I think commadore suffered more from its brand than anything. It was really tough (and ultimately unsuccessful) for them to rebrand from a C64 kids gaming machine to a serious competitor on the desktop with the Amiga. Their previosly established retailers (toysRus, etc) couldn't help them make the switch either.
Ultimately, with this sort of rebranding problem they would have been much better off spending some serious $$ on marketing to shed the C64 stigma in the desktop world.
As a former make junkie I'm definately sold on Ant if for nothing else than the Junit and Manifest file support it offers.
Junit results can be output in XML and run through and XSL transform, then either e-mailed or posted to a website. I do this hourly (mailing the failures) in a nice readable unit test report.
As for building.jar (and variants). Ant has excellent support for putting custom values into manifests such as cvs tags etc.
You could potentially do this in make, but having pre-built libraries for these things and many others makes creating build environments a part-time job instead of a full-time job.
"Nice" and "Toplink" are oxymoronic terms, in my opinion.
Well, unless you NEVER want to migrate away from it. Its like a cancer that invades your entire code-base.
All your code-base are belong to Toplink!
No thanks
Re:Beware of overusing patterns.
on
Design Patterns
·
· Score: 1
On the flip side fitting a problem that exhibits 90% of the charastics of a pattern helps hugely in future maintenance. That is, of course, if the maintainer is well-versed in patterns.
Remember, a pattern is a solution in a context what may look like gratuitous use of patterns may actually be a proper analysis of the context and some amount of compromise for maintenance.
This is a very sketchy and philosophical subject. I've got the underpinnings of a book in progress on this, but so far its more of a philosopy text than a CS book.
Remember when Steve Jobs said that users don't need any more than 128k of memory?
Predictions are very difficult, especially when they involve the future.
In concordance with the other reply. The PPC version was originally ported from OPENSTEP - a NeXTSTEP port to intel. Apple has continued to maintain at least limited functionality of its OS on intel since discontinuing the Rhapsody project. Through the rumour mill I heard the reason Rhapsody was never publicly released was for fears of M$ dropping Office on the Mac.
The currently maintained functionality is realized in the WebObjects runtime on Windows.
I couln't imagine Apple not maintaining a perfectly viable intel-based OS that was up-to-date as late as '99.
FYI: Rhapsody still runs my CVS repository at home.
I remember when SCO first hit the market. You weren't allowed to release software for it without sending your developers to a 2 week developer cert. Needless to say, the company I worked for at the time supported EVERY other flavor of UNIX but SCO. I've been turned off ever since.
You can't bullshit developers? Gimme a break! How else do you explain why MFC &.dll wasn't laughed out of town on a rail? Don't even get me started on the registry. These were marketed by M$ as the next best thing since sliced bread and developers actually bought it.
I would add the IS Survival Guide (can't remember the author). A little dated in process and methodology, but the insights into the industry are dead on the money.
The corporate-speak dictionary is worth the price alone.
Strange you should mention the Apple II reference manual. Its one of two manuals I actually kept from the 80s. The schematics and ROM dumps were too cool to throw out. The other manual I still have is the Microsoft CPM manual - just for grins.
What might be even more interesting than the hardware of yore is the business organization set-up around it.
An interesting job title I remember:
The guy who delivered the punch cards from the engineers to the machine room for execution was called a "Network Connector".
Anyone else remember any goofy titles like that one?
Slashdot Mirror
Those aren't backdoors they're default passwords.
A very different animal, indeed.
I took a look at AspectJ awhile ago and instantly spotted the hidden horror show for implementation: There is no test you can write to assure your aspect was applied appropriately. (at least there wasn't when I looked). Apparently nobody sees this as an issue? The application of aspects is essentially a blind process before compilation occurs. The best you can do is place errors in your aspect and break the compile. I can't imagine turning this tool loose on a development organization.
Seems to me this whole issue is a direct result of MS's tarnished brand. Why bother doing research to find out if this weeks security hole is bogus or not? Microsoft's brand is so coupled with "security compromise" you don't need to prove the case anymore to attain public credibility.
I had a similar problem with a guy at that was placing the mouse on the monitor and following the pointer around with the mouse on the screen. Took about 2 hours on the phone to figure that one out.
Nobody (At least I didn't see it) came up with the oldie-but-goodie:
Terminal Ward.
I think commadore suffered more from its brand than anything. It was really tough (and ultimately unsuccessful) for them to rebrand from a C64 kids gaming machine to a serious competitor on the desktop with the Amiga. Their previosly established retailers (toysRus, etc) couldn't help them make the switch either.
Ultimately, with this sort of rebranding problem they would have been much better off spending some serious $$ on marketing to shed the C64 stigma in the desktop world.
Just my thoughts
Seems to me if apple wants to market the XServer strongly toward mac shops this would be a very nice feature to (re)enable.
As a former make junkie I'm definately sold on Ant if for nothing else than the Junit and Manifest file support it offers.
.jar (and variants). Ant has excellent support for putting custom values into manifests such as cvs tags etc.
Junit results can be output in XML and run through and XSL transform, then either e-mailed or posted to a website. I do this hourly (mailing the failures) in a nice readable unit test report.
As for building
You could potentially do this in make, but having pre-built libraries for these things and many others makes creating build environments a part-time job instead of a full-time job.
"Nice" and "Toplink" are oxymoronic terms, in my opinion.
Well, unless you NEVER want to migrate away from it. Its like a cancer that invades your entire code-base.
All your code-base are belong to Toplink!
No thanks
On the flip side fitting a problem that exhibits 90% of the charastics of a pattern helps hugely in future maintenance. That is, of course, if the maintainer is well-versed in patterns.
Remember, a pattern is a solution in a context what may look like gratuitous use of patterns may actually be a proper analysis of the context and some amount of compromise for maintenance.
This is a very sketchy and philosophical subject. I've got the underpinnings of a book in progress on this, but so far its more of a philosopy text than a CS book.
Maybe some day I'll sort this all out.
That quote is actually relating to the original 128k mac
Remember when Steve Jobs said that users don't need any more than 128k of memory? Predictions are very difficult, especially when they involve the future.
If the plural of octopus is octopi and the plural of server is servi. Is the plural of bus bi? The plural of us must be I.
In concordance with the other reply. The PPC version was originally ported from OPENSTEP - a NeXTSTEP port to intel. Apple has continued to maintain at least limited functionality of its OS on intel since discontinuing the Rhapsody project. Through the rumour mill I heard the reason Rhapsody was never publicly released was for fears of M$ dropping Office on the Mac.
The currently maintained functionality is realized in the WebObjects runtime on Windows.
I couln't imagine Apple not maintaining a perfectly viable intel-based OS that was up-to-date as late as '99.
FYI: Rhapsody still runs my CVS repository at home.
I don't have "net" defined in opts at all and automount works just fine. Strange...
assuming your home directory is on a machine called homehost and your user directory is /home/buba..
/mounts
/Users/buba
/Users/buba. Obviously, this works for other mountpoints and nfs shares in a similar fashion.
From NetinfoManager
Click the padlock to authenticate
Choose
from the directory menu select add directory. change the name of the new directory to homehost:/home/buba
This entry will need the following properties and values:
vfstype nfs
opts nfs
name homehost:/home/buba
dir
This cause homehost:/home/buba to be automounted at
Happy mounting
SCO - ex (as in former, departed, kaput)
I remember when SCO first hit the market. You weren't allowed to release software for it without sending your developers to a 2 week developer cert. Needless to say, the company I worked for at the time supported EVERY other flavor of UNIX but SCO. I've been turned off ever since.
You can't bullshit developers? Gimme a break! How else do you explain why MFC & .dll wasn't laughed out of town on a rail? Don't even get me started on the registry. These were marketed by M$ as the next best thing since sliced bread and developers actually bought it.
Can't bullshit developers... man what a laugh.
I would add the IS Survival Guide (can't remember the author). A little dated in process and methodology, but the insights into the industry are dead on the money.
The corporate-speak dictionary is worth the price alone.
Strange you should mention the Apple II reference manual. Its one of two manuals I actually kept from the 80s. The schematics and ROM dumps were too cool to throw out. The other manual I still have is the Microsoft CPM manual - just for grins.
What might be even more interesting than the hardware of yore is the business organization set-up around it. An interesting job title I remember: The guy who delivered the punch cards from the engineers to the machine room for execution was called a "Network Connector". Anyone else remember any goofy titles like that one?