mailto: blog -at- heyrick -dot- eu

You are not reading my b.log using HTTPS. You can switch to HTTPS by clicking here.

The first eeePC problem

Obviously I have installed software on the eeePC. Most of the programs go into D:\Program Files, but there is still a lot of rubbish that makes its way into the Windows folder...

I downloaded and ran Microsoft Baseline Security Analyzer 2.1 which took an age to scan, but found...

Security assessment: Severe Risk
Computer name: KONNICHIWA\AZUMI
IP address: 192.168.1.69
Security report name: KONNICHIWA - AZUMI (2009-07-18 15-10)
Scan date: 2009/07/18 15:10
Scanned with MBSA version: 2.1.2104.0
Catalog synchronization date: 
Security update catalog: Microsoft Update

  Security Updates Scan Results
	
    Issue:  Windows Security Updates
    Score:  Check failed (critical)
    Result: 25 security updates are missing. 5 service packs or update rollups are missing.

      Security Updates			

        | MS09-010 | Missing | Security Update for Windows XP (KB923561) | Important | 
        | MS06-069 | Missing | Security Update for Flash Player (KB923789) | Critical | 
        | MS08-071 | Missing | Security Update for Windows XP (KB956802) | Critical | 
        | MS08-069 | Missing | Security Update for Windows XP (KB955069) | Critical | 

...
It goes on like that for pages and pages. I had to install Windows Genuine Advantage which I didn't really fancy, but the updater didn't want to continue without it. Then... Well, the updates must have arrived at something in the order of 120Mb - and that's not including the IE8 upgrade, which I disabled as I was expecting a huge file (like MSIE6), but then found out was a much smaller 16.1Mb! I'll get it later...

I went to print a picture and the spool file counted up to 160Mb and then failed. Out of disc space!

The first fault was I had moved my user folder to be D:\RicksFiles, but I forgot to move the TEMP folder. It is now D:\Temp, and the older stuff cleared out.

I also got rid of C:\Windows\SoftwareDistribution\Download. Well, the files are not totally gone, I copied them over to the other machine. I also moved out most of the stuff in C:\Windows\Installer.
My SSD's free space is now just over 1Gb.

Can anybody suggest other things to remove from Windows that while they may be useful, are not actually necessary for the day-to-day functioning of the system?

 

More WiFi fun

Back at the bar yesterday. With their internet access being a freebie (as opposed to €1 per half hour) plus I am using my own machine (useful as I can run my own software and check my own security), I think we'll be going there more often. Okay, it is a lot slower - only 2Mbit instead of about 16, so I set up the big stuff to happen in the background while I was enjoying my meal...

One hint if you plan to leave a file transfer running - the default setting is if the computer is running off battery, go to standby mode after five minutes of 'inactivity'. Best turn this off...

Tell you what, though. Mom likes the Tour de France (as mentioned a year ago) for the scenery. So I looked at the France3 website and eventually found it, and after installing a plug-in, had live streaming video on my computer. It was originally a little rectangle in the webpage, but double-clicking made it full-screen. It fit the eeePC's 16:9 screen perfectly. The quality was a bit fuzzy if you looked carefully, and it hiccuped once in a while, but all in all it was about on par with reading a medium-quality MPEG4 video across 10baseT - I was certainly quite impressed by getting this down a 2mbit ADSL connection (measured at 1.7mbit), plus feeding the data to me by wireless means. Not bad at all.

I don't know the actual bandwidth required for the video - I have my doubts as to whether such a thing would be possible 'at home' if I got 1mbit ADSL, though my lack of hearing anything from Orange (France Telecom) suggests that they aren't getting 1mbit to work - so much for rating the neighbour half a kilometre away as being 18mbit capable!

 

How to delete recursively in Unix...

Now we'd all like rm *.html -r to work. But it won't.

Over to John Bryan who contacted me with an explanation:

Just saw your b*log posting about "rm -rfv *.html" or similar not working has caught me out many times too! Welcome to the 1970's world of UNIX.
 
The reason this happens is that in UNIX, the shell does wildcard expansions for you and the application never gets to see them directly. For instance, with a file directory structure like:
   $ ls -R
   .:
   1.html  2.html  subdir

   ./subdir:
   3.html
Running "ksh -x" starts a new shell instance that prints expanded command lines back to you before it executes them. (In this case "Korn shell", because it would be way too easy to just have one command interpreter in UNIX, instead there are lots and they are all slightly incompatible)....
 
So if I then do:
   $ ls -R *.html
The shell prints the expanded argument list, eg:
   + ls -R 1.html 2.html
Which is what actually gets passed to "ls", which then merrily goes off and lists all the "1.html" and "2.html" files in the directory and all subdirectories...
 
Same thing with "rm".
 
So, in typical UNIX bodge fashion, to get round this you can instead do:
 
   find . -name "*.html" -exec rm -f {} \;
 
Which finds all "*.html" down through subdirectories, and for each file, do "rm -f" (the "{}" is the placeholder for the file found).
(The quotation marks stop the shell expanding the "*". - Note that you can't just do rm -r "*.html" because rm would try and delete a file called "*.html", (because rm knows nothing about wildcards)...
 
Also, the find command is going to call "rm" separately for every single file in question, (which can be fairly slow!)..
 
It would at least have been a little better if you could do:
   find . -name "*.html" -print | xargs rm
Which would only call "rm" once but with the whole list of files to delete, but no, because there are max arg counts an max command line sizes in unix/linux, you can't guarantee that this will work with lots of files!
 
Anyway, that's enough from me, I do like UNIX and use every day for work and so on, but it was never really designed in any way, - it just grew over a period of years, and you can definitely tell!
 
Hope this was useful anyway!
(reproduced with permission, mark-up and styles added for clarity)

Indeed it was John. If such a thing should ever happen again (God forbid!), then we'll know the exact incantation to get rid of a bunch of files in one go. As to the issue regarding Unix utilities being designed to do a specific job well (as in rm deleting a file), it starts to come apart when you have more advanced requirements, such as deleting all of a certain type of file - and in this case all sorts of factors get in the way: rm not understandng wildcards itself, the shell being over-enthusiastic about its wildcard expansion, eventually leading to the bodge of having to 'find' the files to pass to 'rm' one by one.

This will go in my number two slot of why I think Unix sucks. The number one slot, in case you are wondering, is it's the 21st century. We should be so beyond case-sensitive filename systems now... And while some might like the idea of mount points, I grew up with RISC OS which handled multiple filing systems (even as disparate as CDROMs and various harddisc partitions and Econet) with a simple flexible path specifier system. I note that even in XP, I note that I cannot change directory across discs with one command:

  D:\RicksFiles>cd c:\windows
  D:\RicksFiles>
I have indeed selected C:\Windows as the current directory on C:, however I must then:
  D:\RicksFiles>c:
  C:\Windows>
in order to actually go there!
Anyway, RISC OS multiple devices and paths, pretty much the direct extreme opposite to Unix-like mount points (what sucks third place). Never caused me any grief, and I could tell exactly from the path what device, and where on it.

 

Something else RISC OS got right...

Something else I believe RISC OS got right was in replacing file extensions with metadata. In this case, a "file type". Oh sure, it has caused a number of headaches along the way - namely the rest of the world seems to be stuck on file extensions (.doc, .jpeg, .txt...) so RISC OS needs not only a way of transparently supporting translations between the two, but also a place to hide additional metadata for things that are RISC OS specific and won't translate; though things are worse on the 8-bit side of Acorn life as they don't have filetypes but rather a set of load/exec addresses which need to be preserved.
Another problem with RISC OS filetypes is that rather than looking to devise something new, the load/exec system was replaced by a five-byte time stamp which left three bytes for the filetype. 24 bits. Or 4095 different file types, a lot of which are now allocated, in part thanks to some companies claiming loads of filetypes for data resources that don't really need a type back in the early days. It would make sense that a file that the user can double-click on and have something happen should have a filetype. I am not convinced that this is true in all cases... There is further problems with multiple 'similar' files having different filetypes. For my !Teletext software, I support a number of different teletext-like filetypes. I resolved this nonsense with WinTTX which nominally saves the file with a .ttx extension (so a 'type' can be associated with an application), however the loader actually looks at the file to determine what sort it is - thus pretty much all of the RISC OS teletext files that I am aware of can all map to .ttx for WinTTX - and if there is a type that is not supported, given a spec I'll add it. Still, this situation is no different to the old DOS days when people seemed to use .txt and .doc interchangeably (which is fun with Word picking up on all the .docs that aren't really...).

In any case, using metadata for the file leads to a cleaner view of the world. This can be mimicked in Windows if you tell it to hide the file extension, but this leads to the problem of making it really rather difficult to change said extension (i.e. .htm to .html, often arriving at a mess like .html.htm!).

 

Why is the world going who-hoo over XML?

There have been suggestions that "for compatibility", software such as OvationPro and Word (etc) should change to using XML filetypes instead of some form of 'proprietory' format.

Why XML? Why would it not suffice to document your format?

I ask this because to me XML looks like the application version of HTML. Like all the markup with none of the content! This leads to some obvious issues:

  • Unless every possible idea has an XML tag, if you do anything 'different', you will either need to devise a new XML tag or break your software to be compatible with some sort of existing 'benchmark' software. Examples? OvationPro and Word can do a lot of the same things - however OvationPro can 'skew' text. Perhaps not really altogether useful, but if the two programs used XML files, how would this effect be translated? Unless all word processing/DTP software offers all the same functionality, there will still be issues of compatibility and possible data loss when using XML.

  • Why are we taking data that is held in an in-memory structure and converting it to a more-or-less human-readable format? This will then necessitate the step of parsing said data back to computer data, along with the step of validating the data to ensure it is correct (as if we're all using XML, who knows what software might have created the file).
  • And upon parsing - what action do we take upon an unknown tag. Do we ignore it? Do we fault it? If we ignore it and resave without, have we totally messed up the file?
  • Bitfields are common. You can express loads of options in a single computer word - autosave? how often? disable load checking? snap to objects? snap to guidelines? Why are we represting something simple in a load of XML? People complain of application bloat, only now we're looking at data bloat...
  • I assume files will retain their usual extensions - for to make everything .xml would destroy any file association. But if we use existing file extensions, the loader will need to be able to support both 'legacy' (binary) files as well as XML types.
I think OvationPro has got this exactly right. The binary format is not documented, but there is a textual script version. This also is not officially documented, however the format is fairly well known thanks to people playing with it. I wrote an HTML to OvationPro translator and for the OvationPro side, I just knocked up a document with some things (text in bold, colour, centred...) and saw what DDL it generated for it.

So all in all, I don't see how XML will make everything work well. The problem isn't so much a need to arbitrarily change file format. Instead we need more documentation of existing formats and developers/companies more willing to answer requests from people who want to make their products interoperable.

 

Those eight-bitters back in the fast lane

Several nifty things have turned up along the way for the BBC Micro/Master range of computers, proving that despite their age and woeful lack of power when stood next to the latest piece of Vista-compatible landfill, life is still moving along in the land of eight bits.

The first is Sprow, well known in Beeb circles, who has released a network card for Master(Compact) machines. Fitting in place of the Econet card, it will offer 10/100mbit ethernet capabilities. I am not sure exactly what sort of bandwidth is actually possible, the 6502 running flat out can achieve 16mbit/sec, but you really need a load of other accesses (to memory, software/firmware, other hardware, screen memory, etc) to make the system useful. However, even running slower (and remember the ethernet chip will have buffering), the ability to connect a BBC Master to a network of XP machines is, well, quite nice. The only downside I can see is that it replaces the Econet card, making migration or gatewaying problematic. For my own Econet/TCPIP gateway, I planned to use an A5000 with one of each interface, but I never managed to get hold of a copy of !Gateway...
For those interested - http://www.sprow.co.uk/bbc/masternet.htm

Next up, those old harddiscs are starting to wear out, not to mention becoming ever more difficult to replace with compatible types. To help with this, the RetroClinic is offering a CompactFlash based solution. It plugs in, looks and feels like a harddisc, but is smaller, quieter, faster, and more power efficient.


[original image from advert posted on eBay]
Looking at it, it would appear to be a gadget that plugs into the 1MHz bus and delivers an IDE interface. This isn't overly difficult as can be seen by the fairly simple circuitry. More of the smarts of how this works will be in the supplied firmware. CF cards are basically IDE devices, and so the thing hanging off the IDE interface is a convertor to make it CF-friendly.

But wait... hold your breath...

One thing missing from a lot of the Acorn range is USB support, for most RISC OS computers predated it, and serial interfaces were second class to parallel types back in the Beeb's day. Who'd have imagined then that a two wire serial connection could support a bus topology with hot-swapping and speeds that can go up into the hundreds of megabits per second? Well, think that no longer, for RetroClinic has done it again. Not yet released (it is in development), is a Beeb USB expansion - and look at the different connectors on the right for it's a two-way deal here. The Master can connect to USB devices, or it can be connected to and 'seen' as a mass storage device, what with the 1Mb RAM on-board that will eventually act like a sort of disc. Or if you'd prefer a real disc, this astonishing device has IDE on board too. Wow. I wonder if there will be an option of IDE-CF? There are alternatives. If you don't want a slave port, a second host port is possible.
It's called the DataCentre.


[original image from RetroClinic website]

 

Yahoo! mail offline

If you are like me and you want to be able to read your Yahoo! emails offline, it is quite annoying to find them all redirecting you to the Yahoo! website. Thankfully this is fairly simple to work around.
Open the email (the htm file) and look for the part highlighted in the picture below. It changes location from time to time, at the moment it follows all the style stuff.
Select from the "if" to the closing brace, as indicated.
Delete it, then save the file. The email will now open as normal.

Please note that the cookie checked for - in this case "HCdWKBLQ" changes from email to email. Yours will be similar but different. It doesn't matter all the same, delete it.

Actually... you know what? That's crap. That's so much bother - especially if you can't tell JavaScript from CSS having seen neither before. It's a hassle. In the long run, it'd be quicker to know up a program to automate this process...

...which is exactly what I have done!

There is no installer or shortcut, it's just that. Run it.
(you may need my Various Dependencies pack for the VB5 runtime - try it and see)

When you start it, a file selection window will appear asking you to choose which Yahoo! saved mail file to 'tweak'. Once selected, the program will look for the part that needs tweaking, and tweak it as required.
The program creates a copy of the file with "_tweaked" appended to the filename, hence msg_from_keiko.htm will become msg_from_keiko_tweaked.htm - the original is not altered. Nothing else is changed, so references (CSS, images...) will use those of the original file.

When the conversion has been performed, you will be given the option of loading the document into your browser.

 

Your comments:

Please note that while I check this page every so often, I am not able to control what users write; therefore I disclaim all liability for unpleasant and/or infringing and/or defamatory material. Undesired content will be removed as soon as it is noticed. By leaving a comment, you agree not to post material that is illegal or in bad taste, and you should be aware that the time and your IP address are both recorded, should it be necessary to find out who you are. Oh, and don't bother trying to inline HTML. I'm not that stupid! ☺ ADDING COMMENTS DOES NOT WORK IF READING TRANSLATED VERSIONS.
 
You can now follow comment additions with the comment RSS feed. This is distinct from the b.log RSS feed, so you can subscribe to one or both as you wish.

No comments yet...

Add a comment (v0.11) [help?] . . . try the comment feed!
Your name
Your email (optional)
Validation Are you real? Please type 72922 backwards.
Your comment
French flagSpanish flagJapanese flag
Calendar
«   July 2009   »
MonTueWedThuFriSatSun
  1245
6789101112
15161718
202122232425
27283031  

Advent Calendar 2023
(YouTube playlist)

(Felicity? Marte? Find out!)

Last 5 entries

List all b.log entries

Return to the site index

Geekery

Search

Search Rick's b.log!

PS: Don't try to be clever.
It's a simple substring match.

Etc...

Last read at 12:16 on 2024/03/29.

QR code


Valid HTML 4.01 Transitional
Valid CSS
Valid RSS 2.0

 

© 2009 Rick Murray
This web page is licenced for your personal, private, non-commercial use only. No automated processing by advertising systems is permitted.
RIPA notice: No consent is given for interception of page transmission.

 

Have you noticed the watermarks on pictures?
Next entry - 2009/07/26
Return to top of page