If the following item from NTK is true, it all sounds a little suspect from Bigfoot's end. I certainly haven't been told about forwarding limits and the like:
From NTK (http://www.ntk.net) this week:
DESPAIR for remaining BIGFOOT users, whose "free e-mail
forwarding for life" looks to have an expectancy of somewhat
less than five years. "Times have gotten tough for Internet
companies such as ours", the company weasels, and goes on to
say they're spreading that "tough" sensation by limiting
mail forwarding to 25 messages a day. And even then,
subscribers have to provide "demographic information"
before the forwarding restarts. Given that most people
receive about that much mail in "GREAT INKJET DEALS" alone,
you can understand that the company would introduce this
change gently. But, so far, as far as we can tell, they
haven't mailed users, or put the policy change on its
Website - they just stuck it in the autoresponder FAQ at
help@bigfoot.com. How stealthily desperate can you get?
http://www.bigfoot.com/
I've just checked Bigfoot via help@bigfoot.com, and I got:
Our basic Forwarding service is free, but if you wish to use it, you will need to provide us with some demographic information, so that we may send you periodic emails from our partner companies. You will then be able to have up to 25 emails forwarded daily.
I'm constantly surprised that Zsh doesn't receive more attention. I guess the major headache is the lack of accessible documentation which addresses the advantages.
I use it as my default shell everywhere I work, on both NT and Unix. Here's some of the features that I love:
Programmable Completion
You can define completion for different commands in different ways. eg.
ping/telnet completes with the list of machines on my network
man completes with my man pages
kill completes with job numbers
java completes with all my available Java classes (I have to provide a Perl utility to help with this (classfind) though)
Recursive and intelligent wildcarding
ls **/* completes through directories, subdirectories and the like. I never really have to use find anymore
ls **/*(ah-5) finds files accessed within the last 5 hours
ls **/*(L0) finds files of zero length
ls **/*(@) finds symbolic links
etc.
Jerry Peek has quizzed the Zsh mailing lists for info so he can write about it in the next Unix Power Tools. Hopefully Zsh's profile will be raised somewhat.
Of all the people I know using open-source (GNU/Linux) code, I don't know any using Eric's stuff. I don't know any advocating Eric's philosphies, either. I've heard Richard Stallman, and I have disagreements with some of his views, but I know he's considerably more influential. I don't think Eric really delivers anything, whereas Richard usually has something meaningful to say. It's at least tangible enough for me to get a hold of and disagree with, if need be.
Perhaps I'm just overly concerned with his love of guns to take his rants seriously. Or perhaps I'm too British:-)
Liberals are very fond of blabbing on about how third world peoples are exploited in 'sweatshops', and have child labour thrust upon them.
Yup. Including the 'liberals' in these countries who try to set up unions and workers collectives to attempt to improve workers' rights. Who are shut down by the authorities, intimidated by the factory owners etc.
Although these people are making money, I don't think it's right to exploit them by forcing 7 day weeks, 16 hour days, squalid work conditions, no holidays etc.
Read No Logo by Naomi Klein.
I'd urge anyone interested/curious to read "No Logo" by Naomi Klein, which addresses the issues of branding / sweatshops etc. It's a fascinating and informative read.
I've already run into problems with financial applications using 32 bit dates that run out in 2038, unbelievably. With one application I found that we couldn't enter a bond that expired past Jan or Feb 2038 (I worked out the exact date using a quick C program and confirmed this in the application).
So what ? Well, the US treasury issue bonds that redeem in 30 years time. So I suspect the financial industry will see these Y2.038 bugs in 8 years time!
I have to disagree with this. Certainly WAP isn't a substitue for the web, but services tailored for WAP are useful (if in short supply)
eg.
Yahoo's services are now WAP enabled. So I can check my mail, check weather reports, stocks etc. I appreciate that reading mail is not easy, but it's nice to be able to check whether you've received an important mail. I'd prefer to have difficulty reading it than not to even know of its existance.
Multimap's WAP site gives you street by street directions between postcodes (only tried it in London). Excellent for when I don't have my A-Z. Note that this is text only - no graphics.
Railtrack's WAP site gives you timetables for train services. Enter your starting station and end station, and you get the next service available, together with earlier/later options. You don't even need to enter the full names of the stations. It'll work it out and present you with options.
The CurryNet site will give you curry houses in your postcode, with nearby taxi firms, tube stops etc. Useful with site #2 above. Usually by this stage I'm in no state to operate the phone though:-(
My own consultancy WAP site http://www.oops.demon.co.uk/wap/index.wml ! Did I hear 'bandwagon' ?
Most of my work is Java-based, but I can use JPython to script using my Java classes, which makes life very easy. Test harnesses are a doddle, and I can interrogate my class via an interactive prompt
eg. I can never remember the formatting strings for SimpleDateFormat. With JPython I just play around at the prompt until I've got what I want, and then paste the format string into my main project.
I can also write CORBA-enabled scripts, and interrogate my servers interactively.
Much as I love Perl, I don't think Python (or JPython) can be beaten for this!
Perhaps some people view their local culture as being something different than an extension of their in-work geek persona -eg. music/theatre/food/whatever.
Currently I see a lot of moaning about the loss of downloads, and not much addressing of the issues above.
If you don't want to buy a CD, then don't. From my perspective, some of my CDs have given me so much joy that (retrospectively) they're worth 10 times what I paid for them. Others I could happily live without. But that's a risk I take buying a CD. In the UK I'm paying up to £16 in the shops. If I don't want to take that risk then I wait for it to go on sale.
But I'm not being exploited. I think there are many many more people in the world being exploited than music fans, and music artists for that matter. They now have the means to make music and distribute themselves given the web (see the Marillion web site and their initiative for their next album.
On the subject of giving away software, I do that. But the stuff I give away has been stuff that I've built as a consequence of doing other work. I don't buy music that's been generated as an aside.
I'd suggest that Java doesn't enforce type-checking as strongly as people think. eg. if I extract an object from a vector:
Object o = list.get(i);
then to do anything useful I have to cast that object o into the object type I believe it to be. Of course there are various solutions to this (templated Java etc.) but the problem exists in vanilla Java.
That's right. And when you insert an object into a container, it simply checks that it's of type object. And when you extract an object, similar checking is performed.
Consequently I can take my set of bond objects (say), and insert an equity with no problems. I pull my objects out (as type Object), and I have two choices. Either check the type dynamically (using instanceof), or just assume (hope/pray) that they are all really bonds. Of course if I guess wrong I'll get a ClassCastException at run-time.
Templates are the single most important missing feature of Java. I've ended up writing Perl scripts to define versions of Lists, Vectors etc. that enforce type checking by rewriting the method signatures appropriately.
However, there does seem to be a lot of stuff left over from C/C++ days that is sort of questionable. Like the inclusion of structs and enums. The potential performance benefits of structs are intriguing, but losing much of the OOP nature of your code is not attractive. Enums seem to clutter the language needlessly.
Of course the great thing about enumerations is that you have a defined set of values which the compiler knows about. Consequently when used in switch statements the compiler can alert you to missing cases. (useful in instances where you don't want to define sets of objects and use dynamic dispatch on them).
I've just had to can my Newton 2100 after it lost its stylus tracking ability, and I've moved to a Palm Vx. Despite being smaller and the fact that it syncs to loads of stuff (eg. Yahoo address/calendars), I miss the Newton enormously. The Plam uses Graffiti, so it recognises each letter. The Newton used the whole word to perform a lookup, so provided you wrote the first few letters well, it would stand a good job of getting the word right. If it didn't, you'd get a list of other options to select for each word. As Graffiti is character based, you don't get that option. Consequently, taking the teaching time of the Newton into account, I can write into the Newton at a much faster, more accurate rate than on the Palm (entering postcodes/zip codes aside!). Let's hope this technology emerges in another product soon, and I can put my Palm back in the toybox:-)
The reason why I'm pro-XML is that it naturally supports hierarchies. Every time I've written a 'flat' file config, at some stage I've wanted to introduce hierarchies. At that point the parser requires much more work.
Unfortunately I'm somewhat anti-XML for two reasons:
a) I don't believe it's truly human readable, in the same way that, say,/etc/services is. Although you can make a pretty good stab at it.
b) the greatest thing about the conventional flat-file configs is that you can comment stuff out easily. I know you can introduce comments into XML, but commenting configs using # beats using any XML construct hands-down.
Slashdot Mirror
If the following item from NTK is true, it all sounds a little suspect from Bigfoot's end. I certainly haven't been told about forwarding limits and the like:
From NTK (http://www.ntk.net) this week:
DESPAIR for remaining BIGFOOT users, whose "free e-mail
forwarding for life" looks to have an expectancy of somewhat
less than five years. "Times have gotten tough for Internet
companies such as ours", the company weasels, and goes on to
say they're spreading that "tough" sensation by limiting
mail forwarding to 25 messages a day. And even then,
subscribers have to provide "demographic information"
before the forwarding restarts. Given that most people
receive about that much mail in "GREAT INKJET DEALS" alone,
you can understand that the company would introduce this
change gently. But, so far, as far as we can tell, they
haven't mailed users, or put the policy change on its
Website - they just stuck it in the autoresponder FAQ at
help@bigfoot.com. How stealthily desperate can you get?
http://www.bigfoot.com/
I've just checked Bigfoot via help@bigfoot.com, and I got:
Our basic Forwarding service is free, but if you wish to use it, you will need to provide us with some demographic information, so that we may send you periodic emails from our partner companies. You will then be able to have up to 25 emails forwarded daily.
I use it as my default shell everywhere I work, on both NT and Unix. Here's some of the features that I love:
Programmable Completion
You can define completion for different commands in different ways. eg.
ping/telnet completes with the list of machines on my network
man completes with my man pages
kill completes with job numbers
java completes with all my available Java classes (I have to provide a Perl utility to help with this (classfind) though)
Recursive and intelligent wildcarding
ls **/* completes through directories, subdirectories and the like. I never really have to use find anymore
ls **/*(ah-5) finds files accessed within the last 5 hours
ls **/*(L0) finds files of zero length
ls **/*(@) finds symbolic links
etc.
Jerry Peek has quizzed the Zsh mailing lists for info so he can write about it in the next Unix Power Tools. Hopefully Zsh's profile will be raised somewhat.
Of all the people I know using open-source (GNU/Linux) code, I don't know any using Eric's stuff. I don't know any advocating Eric's philosphies, either. I've heard Richard Stallman, and I have disagreements with some of his views, but I know he's considerably more influential. I don't think Eric really delivers anything, whereas Richard usually has something meaningful to say. It's at least tangible enough for me to get a hold of and disagree with, if need be. Perhaps I'm just overly concerned with his love of guns to take his rants seriously. Or perhaps I'm too British :-)
I didn't finish this. Who has time to read 6 paragraphs ?
Liberals are very fond of blabbing on about how third world peoples are exploited in 'sweatshops', and have child labour thrust upon them. Yup. Including the 'liberals' in these countries who try to set up unions and workers collectives to attempt to improve workers' rights. Who are shut down by the authorities, intimidated by the factory owners etc. Although these people are making money, I don't think it's right to exploit them by forcing 7 day weeks, 16 hour days, squalid work conditions, no holidays etc. Read No Logo by Naomi Klein.
I'd urge anyone interested/curious to read "No Logo" by Naomi Klein, which addresses the issues of branding / sweatshops etc. It's a fascinating and informative read.
I've already run into problems with financial applications using 32 bit dates that run out in 2038, unbelievably. With one application I found that we couldn't enter a bond that expired past Jan or Feb 2038 (I worked out the exact date using a quick C program and confirmed this in the application). So what ? Well, the US treasury issue bonds that redeem in 30 years time. So I suspect the financial industry will see these Y2.038 bugs in 8 years time!
eg.
See my related post on this!
Most of my work is Java-based, but I can use JPython to script using my Java classes, which makes life very easy. Test harnesses are a doddle, and I can interrogate my class via an interactive prompt
eg. I can never remember the formatting strings for SimpleDateFormat. With JPython I just play around at the prompt until I've got what I want, and then paste the format string into my main project.
I can also write CORBA-enabled scripts, and interrogate my servers interactively.
Much as I love Perl, I don't think Python (or JPython) can be beaten for this!
Perhaps some people view their local culture as being something different than an extension of their in-work geek persona -eg. music/theatre/food/whatever.
mv ~/**/*(m-7) /mnt/zip
or similar
If you don't want to buy a CD, then don't. From my perspective, some of my CDs have given me so much joy that (retrospectively) they're worth 10 times what I paid for them. Others I could happily live without. But that's a risk I take buying a CD. In the UK I'm paying up to £16 in the shops. If I don't want to take that risk then I wait for it to go on sale.
But I'm not being exploited. I think there are many many more people in the world being exploited than music fans, and music artists for that matter. They now have the means to make music and distribute themselves given the web (see the Marillion web site and their initiative for their next album.
On the subject of giving away software, I do that. But the stuff I give away has been stuff that I've built as a consequence of doing other work. I don't buy music that's been generated as an aside.
Object o = list.get(i);
then to do anything useful I have to cast that object o into the object type I believe it to be. Of course there are various solutions to this (templated Java etc.) but the problem exists in vanilla Java.
...or is Slashdot getting more and more irrelevant these days ?
Agreed. Not worth posting up at all, in fact.
Consequently I can take my set of bond objects (say), and insert an equity with no problems. I pull my objects out (as type Object), and I have two choices. Either check the type dynamically (using instanceof), or just assume (hope/pray) that they are all really bonds. Of course if I guess wrong I'll get a ClassCastException at run-time.
Templates are the single most important missing feature of Java. I've ended up writing Perl scripts to define versions of Lists, Vectors etc. that enforce type checking by rewriting the method signatures appropriately.
Of course the great thing about enumerations is that you have a defined set of values which the compiler knows about. Consequently when used in switch statements the compiler can alert you to missing cases. (useful in instances where you don't want to define sets of objects and use dynamic dispatch on them).
I've just had to can my Newton 2100 after it lost its stylus tracking ability, and I've moved to a Palm Vx. Despite being smaller and the fact that it syncs to loads of stuff (eg. Yahoo address/calendars), I miss the Newton enormously. The Plam uses Graffiti, so it recognises each letter. The Newton used the whole word to perform a lookup, so provided you wrote the first few letters well, it would stand a good job of getting the word right. If it didn't, you'd get a list of other options to select for each word. As Graffiti is character based, you don't get that option. Consequently, taking the teaching time of the Newton into account, I can write into the Newton at a much faster, more accurate rate than on the Palm (entering postcodes/zip codes aside!). Let's hope this technology emerges in another product soon, and I can put my Palm back in the toybox :-)
The reason why I'm pro-XML is that it naturally supports hierarchies. Every time I've written a 'flat' file config, at some stage I've wanted to introduce hierarchies. At that point the parser requires much more work.
/etc/services is. Although you can make a pretty good stab at it.
Unfortunately I'm somewhat anti-XML for two reasons:
a) I don't believe it's truly human readable, in the same way that, say,
b) the greatest thing about the conventional flat-file configs is that you can comment stuff out easily. I know you can introduce comments into XML, but commenting configs using # beats using any XML construct hands-down.