Ask Bloggie

October 21st, 2005

Using perl with RH and apache

Posted by Paul Murphy @ 5:54 am

Categories: Advice

Tags:

Dear Bloggie:

Problem #1: The musicians recently had a camp out for jamming, but mostly eating. Our normal approach is to take a hard copy sign up sheet to a weekly class and obtain volunteers for chili, salad, drinks, etc. It would be much easier if the musicians could simply sign up on their web site. Enter PERL CGI.

Back shortly after the Wright Brothers were flying (I may have been holding the fire bottle) I became a cracker jack FORTRAN and BASIC programmer. I was hopeless with C and C++ (started using profanity around pointers), but really like what I see in PERL and its ability to be used to create interactive web pages. I have installed, written PERL code and run it on my Windows machine.

Could you mention the procedure for installing and using PERL in the Apache - Redhat environment? I’ll be using cgi.pm. If I can get an elementary PERL CGI script to execute from an example web page, I’ll be able to expand that to the food sign-up sheet.

Problem #2: I would like to allow the musicians to provide e-mail feedback on a web page, but when I saw a friends set up was disappointed to see that the feedback received from her script had a large amount of strange characters. You could pick out the content, but it would be nice if it looked more like what was typed as feedback. I hope this makes sense. I’m assuming this was either a feature or limitation of PERL?

Signed:
in search of a pointer

For most people there is no magic procedure to installing PERL for Apache on Linux - it’s just there to do with as you wish. There is one for high volume users - you need to be sure the applicable Apache 2 modules are installed and set up your directory tree carefully; but since you’re not planning on high volume work, you should just go with what’s there by default.

Your apache httpd.conf file will tell you where the cgi directory is:

ScriptAlias /cgi-bin/ "/usr/local/apache2/cgi-bin/"

You can find out where Perl is by using "% which perl" on the command line, or if that doesn’t work try

% find / -name perl -print |& grep -v cannot

or what ever works in the shell of your choice (I use csh)

Put the location info for the perl variant you want to use in the first line of your script:

#!/usr/local/bin/perl
rest_of_your_script

Put the script in $ScriptAlias, set permissions so apache can read and execute the script, and start your HTML form def with:

<FORM ACTION=/cgi-bin/my_script.pl METHOD=POST ENCTYPE=application/x-www-form-urlencoded NAME=my_script>

Now, for question two, I’m guessing that there’s a problem in character conversion on post style submissions. Notice that "urlencoded" above? That’s not Ascii, so you have to convert to a human readable form before doing anything with arriving data.

I believe there’s a slick and generalized solution in CGI.pm but as a practical matter I program in PERL so rarely now that I can never remember the name or syntax needed for the libs and so default to doing simple stuff the simplest possible way - in this by case copying in a bit of script I’ve used since at least the mid ninties:

if ($ENV{’REQUEST_METHOD’} eq ‘POST’) {
read(STDIN, $buffer, $ENV{’CONTENT_LENGTH’});
@pairs = split(/&/, $buffer);
foreach $pair (@pairs){
($name, $value) = split(/=/, $pair);
$value =~ tr/+/ /;
$value =~ s/%([a-fA-F0-9][a-fA-F0-9])/pack("C",hex($1))/eg;
$FORM{$name} = $value;
}

Now everything in $FORM is in the right form…

There’s much better code around, but this works - and you’ll be able to find an equivelent easily enough now that you know what to look for.

On the other hand, you might be better off not trying to re-invent the wheel - have you looked for applicable software, like phorum, that would let you do more of what you want with less effort?

October 20th, 2005

Why Windows Explorer pauses and how to fix it

Posted by George Ou @ 1:48 am

Categories: Advice

Tags:

Hello everyone, I’m George Ou and I normally write for my security corner.  I’ve just come on board here at Ask Bloggie to field questions on PC and Wireless Networking questions.  This is my first reply.

Murph recently received this help letter on sluggish Windows Explorer performance.  Our reader sees very slow performance in Word when browsing for documents using the File Open dialog which is heavily reliant on the Windows Explorer (not to be confused with Internet Explorer).

Our reader sees the following symptoms:

Then the reader writes:

Is there a way to "fix" something so that these long waits can be avoided? I am using XP Pro and when it was first installed (about 1 month ago) every thing was very fast. Since then a gremlin has entered the system. I do regular Norton checks and have had no virus activity. I also check with AdAware and Spybot and have had no significant problems reported in that area.

I am really looking forward to your response and will be watching your new column as it should be both interesting and helpful.

Here is my reply to our reader:

While I’m not sure what you meant on the fifth line by "Left click on ‘Word’" being very slow, I am very familiar with the third line where clicking on the arrow next to "My Documents" causes a severe pause.  I’ll bet this is actually the same behavior you get when you click on "My Computers" on your desktop.

Note:  This problem is so common and annoying that John C. Dvorak actually wrote this article complaining about Windows sluggishness.  While Dvorak is correct in identifying a problem, his assessment and solution was way off and wrong.  By the way John, flash disks are not an elixir and they actually have slower transfer rates than hard drives and Vista actually boots much faster than Windows XP.

The problem is poor design in the Windows File Explorer.  Windows File Explorer attempts to seek volume statistics whenever you open "My Computer" so that it can give you a bird’s eye view of your available resources.  The problem arises when one of those resources, like a CD-ROM drive with a bad disk inside or mapped network drive, is unavailable and forces Windows Explorer to wait endlessly for volume statistics before it times out after a painful 10 to 30 seconds wait time.  Frankly, this is probably one of the most annoying aspects of Windows and contributes to the perception that Windows is sluggish when it doesn’t really need to be.  As time goes on, more dead drive mappings accumulate on the system and makes it seem like your computer is slowing down.  I wish Microsoft would fix this issue because it really makes their flag ship product look bad when people are stuck waiting 10 to 30 seconds every time they come across the "My Computer" folder in Windows Explorer.

The way to work around this is to un-mount all the network drives and make sure there are no bad disks in the CD or DVD drive.  Bad floppy disks are also pretty notorious for causing these problems.  To un-mount all the network drives, you can right-click on the mapped drive and hit "disconnect network drive".  The other thing that seems to help is to clear your most recently accessed documents list.  This was an old Windows 98 trick which had a severe pause whenever something on the recently accessed document list isn’t accessible because it’s on removable media like a floppy disk.  This doesn’t seem to be a problem with Windows XP anymore, but I’m still paranoid about that issue so I still like to clear out that list.  You can do this by right-clicking on "Start" - "Properties" - "Customize" - "Clear".

Note:  Corporate notebooks are notorious for this problem because the corporate domain controller will often mount network drives for you as soon as you log on to the domain.  In this case, un-mounting the network drives won’t keep them from coming back and you’ll need to demand that your Windows NT or Active Directory administrator stops mapping drives for you.  If they refuse, take it over his/her head and escalate the issue and demand this change.  Now before you windows administrators jump all over me and say "what about my standard drive mappings we need", I’d like to remind you of a thing called a shortcut.  You can easily place shortcuts on the desktop pointing to network shares using UNC paths.  UNC comes in the form of "\\servername\share\%username%" which will dynamically point to the actual username folder of the user who invoked the shortcut.  The nice thing about shortcuts is that Windows Explorer doesn’t bother to scan for shortcut resources which avoids the dreaded dead resource wait.  You can even push out the desktop shortcut via login script or Active Directory group policy using a simple batch file or windows script.  The only reason drive mappings were invented was because of legacy DOS applications that didn’t know how to handle UNC paths.  This really isn’t relevant today and there is no reason to use drive mappings especially when dead mappings cause so many problems.  Drop those drive mappings like a bad habit!

Now there is no guarantee that this will fix our reader’s problem, but it is a very good place to start.  If it still doesn’t fix the problem, we’ll need to go in to MSCONFIG and disable all the startup gunk that was placed in there since the clean Windows install.  This not only fixes all sorts of mysterious Windows problems, but it also brings Windows XP boot time to about 30 seconds on a modern PC.  I’ll leave that for another blog.

October 14th, 2005

Bloggie asks, Linux on a Mac?

Posted by Dana Blankenhorn @ 10:26 am

Categories: Advice

Tags:

Dear Bloggie:

My main computer runs Windows XP. I have a second computer that runs MAC OS X 10. I would like to install a version of Linux on the MAC. Is this possible? If so what is the best version of Linux to use as well as easiest to learn.

Dear Skipper:

Since you use Mac OS X I assume you’ve got a Power Mac. 

MkLinux is the best version of Linux for the Macintosh. (That’s their mascot over there — cute, huh?)  Apple actually helped out on the port, working along a French outfit, Silicomp.

There is a problem, however. It’s not yet at Version 1.0, although you can download the pre-release version here .

The best known people in this space, however, are the folks at YellowDog Linux.    I also hear Paul likes Ubuntu, which is an ancient African term meaning "humanity to others." (They also have a new release out today.)

Paul?

You bet Dana - I’ve personally used both Yellowdog and Ubantu and happily recommend both of them. On the other hand, I do have an alternative suggestion for Skipper.

I’d consider going to Linux from Mac OS X a a bit of a downgrade - so why do it? especially when you’ve got an XP box handy for which any Linux would be an upgrade? Bear in mind two things:

1. that it’s all well and good to re-compile your own software, but its sometimes just plain easier to download and install binaries -of which there are lots for the x86 and not so many for PowerPC; and,
2. dual boot sounds a lot better than it is - but the bugs are pretty much out of it for the x86 world: meaning that you’ll be taking no real risk installing Linux as a boot option on that PC, but you will risk causing yourself a lot of grief on the Mac side.

Besides ( -:) ) keeping your Mac on OS X lets you experiment now with Microsoft’s future interface - including the new Office stuff- so why give up a lead on your competition when you don’t have to?

So there you have it, Skipper. Two answers in one. You can run Linux on a Mac, but do you really want to?

Bloggie

October 13th, 2005

Build or buy, Bloggie?

Posted by Dana Blankenhorn @ 8:39 am

Categories: Advice

Tags:

Dear Bloggie,

I am currently in the market for a newer computer. I am debating between either buying or building a new one. I mainly use the one I have now for email, bills, playing solitaire and my grandson’s use it to play games and send messages to their friends.

What is my best choice price wise - to build my own or to buy a new one. (My son will help me build one as he claims that is the cheaper mention)

Thank you,

Esther

Dear Esther:

Ah, yes, the old build or buy question. (I thought this 1978 Heathkit ad, from Dave’s Computer Museum, might make a nice backdrop to our talk today.)

I remember, when I was just a little Bloggie, all those wonderful Heathkit ads. Build your own computer for less, they said.

(Wistful sigh.)

Unfortunately, it’s much cheaper to buy than to build. A visit to your local Fry’s will show you why.

Just go by and collect all the parts you will need to build a simple
computer, of the type you’re describing. (These days that’s a pretty low-end machine.) You’ll need a motherboard, a case (with power supply), a hard drive, a DVD drive, monitor, and software.

Mark down the prices of these parts, then go across the store and check out the prices of the finished machines.

Not much different.

Then consider the cost of your time, even the few hours it might take you to put all this together.

Margins are now squeezed past nothing.

It’s all about retail mark-ups, which are usually twice what the box cost when it came in the door. Do that with one box and $300 becomes $600. Do that with a bunch of boxes, shipped separately, and it can be $700.

Now there is a purpose to that parts department. If you’re building a prototype and need a part, go to Fry’s. If you want to upgrade your motherboard (but everything else works) go to Fry’s. Need memory and your board can handle it? A second hard drive? Fry’s can keep an old computer going for a good low price, for a time.

But build your own commercial machine? Uh-uh.

All that said, I actually got my latest desktop from an outfit that does just what you want done, called Feather Computers. The difference is they’re buying those parts wholesale, then marking them all up to me. The result comes in below what Fry’s charges for its name brand desktops and it works well.

Good luck!

Bloggie

September 28th, 2005

Streaming .avi and MPEG files

Posted by Paul Murphy @ 3:55 am

Categories: Advice

Tags:

Dear Bloggie:

We produce .avi and mpeg formatted video. Currently folks download these and begin to view them after the download completes. We would like to stream them as well as providing downloads. Besides a Real server, is there an efficient, cost effective way to do this?

Signed,
Anyone from anywhere

Dear Anyone:

Yes.

How cheap, how fast, and how efficient depends on your volume.

MacOS X comes with a streaming server capable of handling all the major formats so if this is something you don’t do a lot of, get an older Mac with lots of RAM and you’ll soon be in business.

Now if you do more of it, but not as much as say a baseball franchise, consider MacOS X on an X-server with an X-RAID. These servers are fast and cheap, you don’t face the Microsoft per client licensing issues, and set up is pretty much out of the box. The X-RAID is wonderful for streaming large videos because it more than makes up for being relatively slow to set up a simple file transfer, by being extremely good at keeping up with streaming requests once the flow starts. Think of it as a diesel and you’ll get the idea: tough as nails and slow to get going, but unstoppable while in motion.

If you get really big the right answer is Sun -but they charge you for the server license and it takes specialized expertise to get it working right. What’s going on there is that this is one of Sun’s little secrets - you know, goodies the marketing goofs never tell anyone about in case, gee, wouldn’t that be terrible, people actually bought the product.

I discovered this a few years ago while researching something similar for Sam Hiser -co-author with Tom Adelstein of Exploring the JDS Linux Desktop and, at the time, trying to get something called reelamerica.com going. ReelAmerica was going to stream out thousands of small videos concurrently and so needed something very agile - and it turned out that not only does Sun have an appropriate streaming server, but it uses custom instructions embedded in the SParchitecture to do it. Duh? who knew? Not me, but apparently quite a lot of buffos in agencies with acronyms for names and, today, major league baseball.

So the general answer is "Yes" but the specifics are going to depend on your volume.

September 27th, 2005

Curious about batteries

Posted by Dana Blankenhorn @ 5:49 am

Categories: Advice

Tags:

Bloggie…

This is more of a tech question than a computer question.  It is related to  batteries?  I have been trying to find a rechargable battery that will work on par with regular alkaline batteries and maintain affordability.  I came across Juice batteries online and tried them out only to find that they meet those needs.  My question is…do you think that the rechargable battery market will flourish with the increase in high energy consumption gadgets (pda’s, mp3 players, digital cameras, etc…)?

Thanks,

Curious in Canton, MI

Dear Curious:

Like you I use rechargeables whenever I can. But one of the most interesting new technologies in the reusable portable power market are fuel cells.

Two years ago, NEC brought out a fuel cell, powered by a half pint of methanol, which ran 5 hours between refills. At the time the company promised that "within two years" models would be available offering 40 hours of life between charges.

A year later all NEC could offer was 10 hours of life on a fuel cell charge-up, with an underpowered machine.

Last week NEC was pushing what it called an Organic Radical Battery (ORB) and an H-P spokesman was quoted as saying fuel cells are "nowhere near prime time" for consumer products.
Yet-another workshop on the technology took place this past month.

Those of us who want to go cross country with our laptops are left with high-priced batteries, like the Energizer e2, that release their charge more slowly than conventional batteries, but are still based on chemical technologies. Target is selling a 6 volt version for digital cameras at $11/each. The AA versions are available in 4-packs from $15-20 per pack. (The package is pictured to the right, from C|Net.)

Bottom line? There’s still some research to do, but we are getting close to some breakthroughs. And with folks like you on the demand side, there’s plenty of incentive out there for people to make those breakthroughs.

Bloggie

September 20th, 2005

Baffled on backup

Posted by Dana Blankenhorn @ 5:48 am

Categories: Advice

Tags:

Dear Bloggie:

 I have an old, old system that I need to do a hard drive clean sweep with;  however, I need to keep the operating system intact (windows 98) or else I won’t be able to run it. It has all of about 3 GB in the hard drive. Is there a way to wipe the drive clean without losing the operating system. This is an old Aspire Acer 6100. (ancient system) I have almost no backup software so if I get into a jam thats it. It’ll be difficult to get out of. Can you help me?

Thanks,

Baffled in Byron Center, MI

Dear Baffled:

Let’s have Paul take the first look at this one, since I’ve lost more data than Rosemary Woods.

Paul?

I’ve been in that position - where you just don’t want to touch the thing because it might collapse but you know you have to. So don’t even think about  touching that disk until you have a tested, working, backup - even if that means putting a second disk in the machine or backing up to a friend’s laptop.

Backup’s first, recovery testing second, futzing third.

Dana?

Excellent points. Once we’re crystal clear on the need for backup, how can you get one that fits your budget?

Personally I use a program called PC Relocator from Eisenworld whenever I get a new PC. It costs about $50. It moves everything and anything, including the operating system. If you have another place to put this stuff (like a friend’s laptop) it’s an easy solution.

But you say you don’t have another computer to move your old OS to?

Yes, you do. It’s called GMail. Copy your operating to your new GMail account as an attachment, along with anything else you need to keep.

You can also get shrink-wrap copies of Windows 98 at eBay so if you want to just wipe your drive and start over that is an option.

If you’re really hard up I have a copy of Windows 1.0 from 1986, hand-signed (in the manual) by Bill Gates himself. But it’s on floppies. I think they’re 5.25 inch floppies.

The point is computing has changed in 20 years, and it continues to change rapidly. Three gigabytes isn’t a lot of storage these days. There are many places you can find it. So long as the hard drive’s problem doesn’t prevent you from  moving things off it, you should be OK.

Just remember Paul’s rule. Backup, then recovery testing, and futzing last.

Bloggie

September 19th, 2005

The wonderful land of Ogg

Posted by Dana Blankenhorn @ 6:48 am

Categories: Advice

Tags:

Hello Bloggie:

I have a question in respects to music files on a website. I have an MP3 song which I would to incorporate on one of my webpages and was wondering  to which format I should convert it for it to be the smallest. I know midi files are small but this is a song by an artist and it is 2.82MB right
now.

I can add it to my page but I am afraid it might be too big. Is there a format that would make it smaller yet still play properly on my webpage?

Thanks

Signed

Francine the Grand Explorer

Dear Francine:

Ogg.

Excuse me, I hear you cry?

Ogg. The answer to your question is Ogg. Ogg is a patent free, multimedia bitstream container that can handle audio or video.

Since you are using MP3, you already have a codec that accepts some loss of signal quality. It’s lossy, in other words. That train has left the station, so why not look at Ogg?

You can tell a file compressed using an Ogg codec because it has the file
extension .ogg.

Vorbis offers Ogg codecs for a range of operating systems. Vorbis hosts a list of third party software using its codec. Here is their FAQ or, if you prefer, the developer version. They claim superior performance to MP3 or the Windows Media Player, but your mileage will vary.

Now I do take your criticism about the file size seriously as well, and fortunately many companies have been working on improvements to codecs that will decrease file size. I interviewed one such gentleman, Richard Wigard, recently for another blog and, while he will have a proprietary solution, he hopes to get it to market next year.

Good luck.

Bloggie

September 16th, 2005

UDDI , JES, and what, exactly?

Posted by Paul Murphy @ 4:36 pm

Categories: Advice

Tags:

Dear Bloggie:

Two questions: How does identity (ideally single sign on) work in Solaris (mainly 10)? We can’t figure out which 3-10 pieces of Sun’s JES we need to do single sign on. LDAP I get but what else do we need?

What the hell good is a UDDI server? I see it as roughly equivalent to a root DNS server (the world won’t need a lot of them) but it seems to me you need to know most of what you’d get from it in order to find what you need (the web service you want) anyway.

Signed:

Over caffinated
in javaland

I’m confused too, and I’ve installed several versions of Sun’s identity management stuff.

On the positive side, we can deal with the UDDI question fairly easily. Back in the late ninties there were people who thought there was money to be made making program widgets available for use across the web - use somebody’s netbean, send them six cents, kind of thing. Naturally that needed a services directory and so UDDI (Universal Description, Discovery, and Integration) was born.

That died, of course, but a few years later Microsoft thought it was going to turn XML into "a programming language for the web" and UDDI got a bit of boost. That hasn’t panned out either, so UDDI has more recently been touted for use in an identity management framework.

Sun adopted it early on because some of its Project liberty partners thought Microsoft meant it, and now UDDI forms a standard part of virtually every inter-operability suite out there. There are situations where the standard can be useful - if, for example, you wanted to implement a "services oriented architecture" (a buzz phrase continuing the generalization from Unix daemon to Microsoft server) the UDDI protocol can work well with basic identity management to point user PCs at the right sockets without further sign-on formalities.

Part one of your question is a lot harder to answer in large part because I find single sign on solutions inherently very confusing. Basically you have two choices with Solaris 10 - implement the most recent version of the traditional single sign-on within a community of trusted machines (including containers) or implement a single identity management scheme.

The former approach, now based on kerberous token passing between machines and an LDAP database, is the logical successor to the original Sun yellowpages (NIS/NIS+) solution. It’s still clutzy (but at least it’s not federated identity!) and it’s still fairly easy to implement - just follow through the manual ( System Administration Guide: Naming and Directory Services (DNS, NIS, and LDAP) or check bigadmin for a step by step guide that fits your environment.

If, however, you want a full blown identity management scheme that extends beyond your business walls to control transactions from outside, I think what you need is the Java identity management suite: Sun Java System Identity Manager, Sun Java System Access Manager, and Sun Java System Directory Server (Enterprise Edition).

I say "I think" because I’ve only used this stuff after loading the entire Sun Java Enterprise System - something I’d suggest you do too because it’s a lot simpler and can be cheaper.

It’s cheaper because Sun has an enterprise licensing deal that is so well defined it can mean almost anything your salesman wants it to.

It’s simpler because the install process puts everything in "the right place," with the right ownerships and permissions according to the defaults assumed in the code - so you don’t have to guess which file, which version name, which oddball permissions, or which unique file ownership expectations are causing your problems.

September 14th, 2005

'Lightweight' Linux?

Posted by Paul Murphy @ 11:48 am

Categories: Advice

Tags:

Dear Bloggie:

I refurb older, donated computer models for senior citizens. Most of the units are Pl with 64 or 128 MB of RAM. I need an OS that won’t hog the memory. These computers are used primarily for Internet access, email and basic games. We only have dial up available so a simplified browser would be nice too. And did I mention, they need to be free?

Signed: Nick in Hicktown

Dear Nick:

First, lets assume that these machines get to be used in one of two situations: at individual homes where they’re pretty much personal machines and in group homes where they’re intended to be shared.

In the personal setting what you need is going to depend a bit on what you have and a bit more on the particular needs of the individual. In most cases, for example, middle aged and older eyes have trouble working with even 1024 x 768 on 15 inch screens. Spread those dots out a bit, and the situation gets better but what this means is most of the time any nickels you’ve got to put into these machines have to go for bigger screens and matching controllers, not Linux or applications.

There’s no problem with big screens and bigger fonts for "lightweight" Linux but there are some issues with other special needs ranging from oversized keyboards and mice to touchscreens and text to speech converters. It’s the devil’s advice, we know, but some of the machines you get in will have legal licenses for NT 4.0 workstation and, on small machines with older peripherals, this stuff works about as well as anything - so maybe reserve your 128MB machines and those licenses for people with applicable handicaps and bite the bullet on sometimes supporting Windows.

For the more generic home user, I’d suggest standardizing on one "lightweight" distribution like Vector Linux. You can use these to make custom install CDs to leave with the machines that, by default, install your choice of windowing manager, browser, larger fonts, and email client.

For example vectorlinux supports both Fluxbox and ICEwm and the KDE office suite along with both Firefox (for the 128MB units) and Dillo (for the 64MB crowd).

Vectorlinux isn’t the only choice and I’m not saying it’s any better than the others, but I am saying you should pick one and stick with it because eventually you’ll stop doing this - and whoever takes over will have an easier time (meaning that your clients will have an easier time) if you standardized early on.

You can do this for group homes too, but I suspect you’ll be better off installing one more or less up to date Linux machine on a faster connection even if that costs money, and then adopting the linux terminal server technologies to run your user workstations as network clients. That will let you attach special purpose hardware (including an NT 4.0 box or two to share the connection via your proxy server) to some user stations, while stealing some memory from them to augment the in-home individual units.

Dana ?

One word: Novell.

Novell has made a lot of its dedication to desktop Linux.

Version 9 of its software includes what it calls a full "productivity suite," which is a generic term for Office (just like photocopy is a generic for Xerox).

The system requirements are (dare I say it) primitive. When was the last time you saw a 266 MHz machine required? Just last month I bought a 2.4 GHz box for about $500.

The base price of just $35 (complete with documentation) is cheap, and even the "upgrade guarantee" and "support" prices are pretty low.

Now, I’ve written often on my Open Source blog expressing skepticism of Novell’s desktop Linux drum-beating. That’s mainly based on its past history, when the company was based in Utah. These days it’s based in Massachusetts (kind of like Governor Romney, who ran the Salt Lake City Olympics before being elected the Bay State’s chief executive).

But what’s at risk here? At these prices, it’s worth a shot.