Main

Bits Archives

April 22, 2007

Friends Don't Let Friends Buy PCs

Last week a friend asked me to help him buy a computer. Instantly, I suggested that he buy a MacBook. I always do this. Call me selfish, but I prefer going to the beach than spending countless hours diagnosing, recovering or teaching my friends the vagaries of a clumsy-at-best operating system.

Alas, my friend didn't have enough cash to buy the MacBook so we purchased an HP dv2-something or other. A reasonable machine. It was bundled with Vista

(cut, interior, friend's house, several frustrating hours later)

Ok - I now have this sinking feeling. What takes 10 minutes to do on the Mac (start and register the machine) has now taken about 2 hours on the Vista box. I was amused when the while-you-wait slideshow mentioned something like: "Vista gives you more time for the things you enjoy". At the bottom of the screen, we were advised to: "Please wait."

Then there's that annoying security checking all-the-freaking-time.

Already - I confess - I've learned. It's the last time I'm going to be an accomplice to a friend buying a PC. (Unless of course, they're going to return the OEM OS and install Linux).

Back to the beach.

Useful Links:


Technorati Tags:
,


May 27, 2007

A New Law For Operating System Selection

I am proposing a new law for Operating System Selection.

Yet another friend asked me at the pub the other day if I would help her buy a new computer. We talked about it for a minute or two while I bandied around venerable selection criteria - such as usability, software requirements, viruses, price, the illuminated Apple logo on the top of the clamshell and so on. My friend was somewhat nonplussed.

It occurred to me that there had to be a much more didactic means of choosing a suitable OS. A selection criteria which would make all others pale into insignificance. Furthermore - the criteria must be technically sound, beyond questioning of the layman and free from FUD, marketing hype and OS wars.

So. Applying Occam's Razor a vast number of OS characteristics using the selection criteria above I arrived at a profoundly simple solution.

I present Campbell's Law Of Choosing An Operating System:

When selecting an Operating System, it shall be bundled with Vi. [But not in an "I can't uninstall it" antitrust kind of way.]

Straightforward, really.

Speaking of Vi - do you ever get in that really annoying situation where you have been using Vi for days, and then switch to Word or something similar in order to type a letter of complaint to your landlord who can't read Courier to find you yourself typing:

iDear Sir,<enter>Please replace my missing front door immediately.<esc>:y100p etc etc etc

It's annoying to say the least - that Word et al don't intercept Vi commands and simply carry out the actions. With accessibility the big issue for web sites and software these days accepting Vi commands should be part and parcel of every application. Rob Sinclair please take note - accessibility is for everyone, including Unix users.

I digress.

Now selecting an Operating System is easy. You go to the store. Pick a machine bundled with Vi. Buy it. Live happily ever after. Empirical evidence to date suggests that I am overwhelmingly more satisfied if you decide according to my law of OS Selection. I contend that this proves my law. The results are repeatable. Prove me wrong, if you dare.

In the event that you still can't decide, use Campbell's Law Of Choosing A New Computer:

When selecting a new computer, the machine shall have an illuminated apple on it.

Back to the beach.:wq


Technorati Tags:
,


June 15, 2007

Safari or Hell Ride?

I much prefer working on a Mac to a Windows machine. Mac OS X Leopard was previewed at WWDC 2007 (World Wide Developers' Conference) and it has some great new features - which makes the view through Windows seem less like a vista and more like a sort of blurry beer-goggled view of the gutter. Steve Job's keynote is always good viewing - with some humor and also nicely balanced by the fact that not all new features are there for the sake of "productivity" - some are there just for fun.

The announcement that Safari was ported to Windows and the following rather uninspiring demonstration was, quite frankly, a waste of ten minutes of my life.

Why?

Because Safari blows.

It surprises me no end that Apple are entering into a market behind the eight ball - that is - with such a featureless, buggy and apparently insecure product. It's failings are well noted on the web, but a quick look at Apple's download page for Safari shows an astonishing list of features "Coming Soon" which should be absolutely baseline: international text, cookie management, FTP directory listings and so on - all basic features which other browsers have had for a long, long time. This will certainly do nothing to entice PC users to start using Macs - on the contrary I suspect.

To make matters worse, Apple intend to bundle Safari with iTunes. Actually - to all manufacturers: When will manufacturers desist from this idiocy? Like WGA with Security Updates, all the HP drivel-ware that you get when you simply want to install printer drivers, and iTunes when all you wanted was QuickTime? It's as insidious as web sites that insist on resizing my browser window. I want to decide what is installed on my computer. Not you. Not anyone else. Please stop.

Back to the beach.


Technorati Tags:


August 1, 2007

ECMA-376 OOXML... And That Hex Dump From MS

I have been doing a large amount of reading on ODF vs OOXML over the last few days - particularly as the verdicts of various countries' standards bodies become finalised. There a quite a number of publicly available analyses of the ECMA-376 Office Open XML Spec available for reading on the internet - comments on Groklaw and the British DIS 29500 Wiki are particularly comprehensive.

Several thoughts strike me as I read though the wealth of material:

Where are the safeguards that prevent interference in standards organisations' decision making processes? Incits would appear to have been stacked over the last few months to create a favourable decision for the proposer of the standard. Surely the composition of voting rights should be decided at the outset of the consideration of a particular proposal. The least of the problems is that the composition of voting members becomes unrepresentative. Other concerns are that a voting member who joins in the last few days prior to an actual vote has not undergone the same exercise in diligence of the proposed standard as the other members.

Furthermore, in the case of documented interference (see www.noooxml.org) there needs to be some mechanism of automatically registering a "no" or "yes" vote, for the standards body in the country that has been interfered with, to counter the interference. This cannot be at the discretion of the standards body of the country involved since their vote is possibly tainted. This effectively immediately removes the sting from any attempted interference.

It would seem that the standards processes are ill-equipped, possibly naïve, for dealing with the intense lobbying and influence of commercial interests. Standards and Standards processes must have unyielding integrity.

Secondly: What the hell were ECMA thinking? They should be downright embarrassed. Given the lack of quality, lack of clarity and lack of technical integrity of the document that they have proposed as an ISO standard, clearly ECMA have failed in their duty at being a trusted source for candidate standards. ECMA should no longer have right to submit a proposed standard for fast tracking.

Thirdly - and this is more a personal note - the countries which I am closely related to - Australia and Qatar - are distressingly quiet on this issue. The Qatar Government is an entrenched Microsoft shop. It's IT environment is immature and somewhat subservient - so I guess it's not much of a surprise that not a wrinkle has emerged from here. Australia's response which amounted to "we're not completely happy with this proposed specification" is insipid in the extreme.

Lastly: Is ECMA-376 indicative of the quality of work that is carried out within Microsoft? The poor workmanship of the ECMA standard speaks volumes about the quality of processes, thinking and product created by Microsoft. It serves to demonstrate yet again that Openness (Source and Standards) yield quality whilst closedness leaves you with a steaming pile of crap. After this demonstration we get a deeper understanding of why Microsoft do, indeed, have reason to fear Open Standards and Open Source.

Back to the beach.


Technorati Tags:
,


October 1, 2007

Escape Firefox's Overflowing AutoFill

Here's one little feature of Firefox that falls into the category of "it's pretty good - but it would be so much better if it could read my mind". AutoFill needs a quick and easy way to manage it's entries - such as About:AutoFill which would be a page allowing you to edit/add/remove various entries. As others have mentioned, after a while AutoFill gets too many entries to be useful.

It turns out the less than obvious Shift + Delete removes entries if you are focussed on them. This was recently pointed out courtesy of Winds Of Change.Net and MacLive.NET.

While we're on the subject of Firefox features - "I want a Firefox Extension to" lists loads of extensions so you can go crazy.

By the way, if you don't use Firefox as your browser already, click on the button below to make the switch:

Firefox 2

February 6, 2008

Leopard Login Screen Picture... UGH

Been meaning for such a long time to work out how to remove this (uncharacteristically for Apple) naff picture from my login screen:

DefaultDesktop.jpg

Needless to say, it's not really my cup-of-tea. I'm surprised Apple didn't use a photo of their Dear Leader - although the actual login window would probably have covered Steve's face.

I hope that Tinkertool add the ability to change this (plus perhaps a few more Dock configuration bits and pieces would go astray).

In the mean time, if you want to change the picture, you need to go under the bonnet and replace this file:

/System/Library/CoreServices/DefaultDesktop.jpg

with the image file that you'd prefer.

There's a longer description on how to do it here.

February 7, 2008

Straightforward DDNS Configuration On Linux

I've just been playing around with DDNS on Fedora again. I do this every few months on a new box so this post is as much an aide mémoire for myself as much as anything, however, others might find it useful. The DHCP DNS interaction on linux seems to be quite poorly Howto'd despite best intentions.

I'm working with an FC8 box. Here's the highlights (I mean, the description is deliberately terse) of the configuration that you need to make Secure DDNS work:

Make An Update Key

Fedora comes with rndc.key supplied. I suggest not using a well known generic key. Generate your own like so:


dnssec-keygen -a HMAC-MD5 -b 512 -n HOST somehost.somedomain.com

This will create your public and private keys. Copy the .key into the update key file you want to use (I use update.key)

The generation process may appear to hang due to a lack of entropy - so if this is the case, open a window and start typing the World's Greatest Novel in another window on the server.

The Key File Format

It should look something like this:


key somehost.somedomain.com. {
algorithm HMAC-MD5;
secret "Nd2+awholelotofseeminglyrandomstuff /moreseeeminglyrandomstuff==";
};

SELinux and Keyfiles

If you're using SELinux, the context of the key file should, for dhcpd, should be:


[etc]# ls -Z update.key
-rw-r----- root named system_u:object_r:dnssec_t:s0 update.key

SELinux will complain if it's not set to this - so use chcon to fix it if necessary.
dhcpd.conf
Important parts are these:

ddns-domainname "somedomain.com" ;
ddns-update-style interim ;
ddns-ttl 86400;


key somehost.somedomain.com. {
algorithm HMAC-MD5 ;
secret ##$$##$$##$$ put your public key data here ;
}


zone somedomain.com. {
primary 127.0.0.1 ;
key somehost.somedomain.com. ;
}


zone 1.168.192.in-addr.arpa. {
primary 127.0.0.1 ;
key somehost.somedomain.com. ;
}

In the subnet section:

subnet 192.168.1.0 netmask 255.255.255.0
{
...
ignore-client-updates ;
ddns-updates on ;
...
}

Now, the named.conf file:

In each zone that you want to allow updates:


zone "somedomain.com" IN {
...
allow-update {
key somehost.somedomain.com. ;
}
....
};

or, if you prefer more control

zone "somedomain.com" IN {
...
update-policy
{
grant somehost.somedomain.com. subdomain somedomain.com. TXT ;
}
...
};

and do something similar in the reverse lookup zone and

include "/etc/update.key";

That's it.

Config testing can be done with:

  named-checkconf
  service dhcpd configtest
Which will show you if you have any issues with your configs before restarting (they are noisy on error only).

I'm dubious about the results from configtest as you'll see later.

Test With nsupdate

Here's a completely uncommented example. Implies that the configs are fine and you've restarted at least named and checked /var/log/messages for anything additional:


[etc]# nsupdate -k Ksomehost.somedomain.com.+157+08869.private
> server 127.0.0.1
> zone somedomain.com.
> update add newhost.somedomain.com. 86400 IN A 192.168.1.1
> show


Outgoing update query:
;; ->>HEADER<<- opcode: UPDATE, status: NOERROR, id: 0
;; flags: ; ZONE: 0, PREREQ: 0, UPDATE: 0, ADDITIONAL: 0
;; ZONE SECTION:
;somedomain.com. IN SOA
;; UPDATE SECTION:
newhost.somedomain.com. 86400 IN A 192.168.1.1


> send
> quit


[etc]# ping newhost
PING newhost.somedomain.com (192.168.1.1) 56(84) bytes of data.
64 bytes from somehost (192.168.1.1): icmp_seq=1 ttl=64 time=0.053 ms
64 bytes from somehost (192.168.1.1): icmp_seq=2 ttl=64 time=0.029 ms
--- newhost.somedomain.com ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 999ms
rtt min/avg/max/mdev = 0.029/0.041/0.053/0.012 ms

Use update delete newhost.somedomain.com. A to remove the entry.

Key File Strangeness / Incompatibility

Unless I'm mistaken, it seems that named and dhcpd prefer the update.key file to be in slightly different format. named needs quotation marks surrounding the key or else it gives this error:


[etc]# named-checkconf
/etc/update.key:3: missing ';' before '/'
/etc/update.key:3: '}' expected near '/'

However, dhcpd complains about the quotes:

Feb 7 somehost dhcpd: /etc/update.key line 3: invalid base64 character 32.
Feb 7 somehost dhcpd: #011secret "xxyyxxxyxyxyxyxyx /xixixkxxjxjx
Feb 7 somehost dhcpd: ^
Feb 7 somehost dhcpd: /etc/update.key line 4: expecting a parameter or declaration
Feb 7 somehost dhcpd: };
Feb 7 somehost dhcpd: ^
Feb 7 somehost dhcpd: /etc/dhcpd.conf line 24: /etc/update.key: bad parse.
Feb 7 somehost dhcpd: include "/etc/update.key"
Feb 7 somehost dhcpd: ^
Feb 7 somehost dhcpd: Configuration file errors encountered -- exiting

Strangely, this causes dhcpd to fail to start (which can be seen in logs) but it is not reported by config check.

I believe that dhcpd is at fault here - the config file should allow the quotation marks.

Versions

I'm using FC8 with Bind 9.5.0-23 and dhcp 3.0.6-12.

Some References

This article is great - very complete - even though it is now quite old:
my newbie DDNS-HOWTO http://www.lugs.org.sg/pipermail/slugnet/2003-January/002848.html

The following articles are great, but they don't have the dhcp config:
nsupdate: Painless Dynamic DNS http://linux.yyz.us/nsupdate/
Painless DDNS part 2: the server http://linux.yyz.us/dns/ddns-server.html

July 11, 2008

Adobe have lost the plot...

After trying to patch my copy of Adobe Reader to 8.1.2 several times - unsuccessfully - I then tried to install Adobe Reader 9.0. This is the result:

Acrobat Crapola.png

A fabulously descriptive error message - wouldn't you agree?

Continue reading "Adobe have lost the plot..." »

August 6, 2008

Why doesn't HP make software that works properly?

I have just fought with a Windows XP machine to try to get HP Scanner software to install properly. I have installed the software 3 times - twice from the disk provided and once downloaded from HP. I have rebooted the machine probably over 10 times and plug and unplugged the scanner probably about the same number of times.

The result is that only the Twain driver is working. None of the actual HP Software for the scanner works. For a non Administrative user, some of the software just gives an error dialogue when you try to launch it. For all users, the HP Solution Center [sic] software doesn't launch at at. So much for a "Solution Center".

Continue reading "Why doesn't HP make software that works properly?" »

August 27, 2008

Facebook "Warning! Your account could be disabled" - Well I'm Confused...

Just logged onto Facebook and inexplicably received this message:

FB Warning.png

It's confusing and nondescript. As a user who logs in daily - sends a few birthday greetings, gets a message or two a week and pokes a few friends overseas now and then (basically - keeping up with my friends!) I'm at a complete loss as to why I should receive this.

Continue reading "Facebook "Warning! Your account could be disabled" - Well I'm Confused..." »