Maybe it’s just me but it seems that creating attributes or some classification meta data on an ecommerce site is only a good idea IF YOU CORRECTLY TAG the items. Else, it produces some head scratching.

Example this morning: I follow a link on facebook to “winter dresses” (I rarely do that but it caught my eyes and I do need more dresses with long sleeves). It puts me on the “short day dresses” category:

So far so good, right, but these are kind of boring so I decide to look at the “long day dresses” category:

I must lack the subtle understanding of the difference…

On a more serious note, it does undermine my trust both in the site itself (not that I had a lot, I don’t know that business since I just followed a random ad) and its ability to help me find what would best suit me…

I have been cleaning my bookmarks and I notice I have a whole set of very simple web based tools I have which I think can be of use to everyone one day or another…

Text manipulation - Textopus http://textop.us/

webforms for just about every text transformation or conversion you might need - without having to write a script :D
“everything from sorting lines to advanced encryption and hashing”

online file conversion http://zamzar.com/ and http://www.pdfonline.com/convert-pdf/

When someone sends you a format you cant open… (in my case, something like microsoft publisher, or a vector file drawing at work)

what is its color http://whatsitscolor.com/

Get the main colors from an image - similar to what you can get on colorlovers.com

Visibility check  http://www.vischeck.com/vischeck/

Vischeck is a way of showing you what things look like to someone who is color blind. You can try Vischeck online- either run Vischeck on your own image files or run Vischeck on a web page. You can also download programs to let you run it on your own computer.

Timestamp converter http://www.4webhelp.net/us/timestamp.php

(geeky) when you’re not on unix and you need to check a date, perhaps from a logfile entry

Recipe measurement converter http://www.ez-calculators.com/measurement-conversion-calculator.htm

This site has many other calculators and no ads…

File extension search http://file-extension.net/seeker/

what’s that file for???

password generator http://www.pctools.com/guides/password/ and http://www.fourmilab.ch/onetime/otpjs.html

Very handy if you’re creating accounts for people

DNS dig http://www.analogx.com/contents/dnsdig.htm and subnet calculator http://www.analogx.com/contents/cidr.htm

(geeky) check domain name records and test DNS servers, and figure out whether /24 really does what you think it does…

connection tests http://www.measurementlab.net/measurement-lab-tools

what is your ISP doing to your pipes…

The mininote can be a pain on many distributions - because it is via based, with the via C7 CPU and a via Chrome graphics card. Also it has a broadcom wifi card which seems to not always be recognised in many distributions, often the distribution’s own b43 drivers did not work on my mininote. SO after lots of trial and error I found a small number which did work. First one, SliTaz

I found out that unless I got certain steps in exactly the right order, things did not work out with the install, or later with the wireless - putting it all up as quicknotes for now, will add screenshots and more explanations later

* Download slitaz ISOs from http://slitaz.org/en/get/

I have used here a recent (April/May/July 2009) version of the “cooking” version, because in early testing the stable version just would not let the wireless work. In the past I have had to manually compile some of the drivers, but it seems the current release has a better package for the broadcom drivers, which means there might be no need for manual compilation as I have had to do in the past. Yay!

IMPORTANT: if you do not have a wired internet connection, download the “packages” version of the cooking (over 1Gb) distribution so you can install the necessary things to make the wireless work

get and run the unetbootin tool

UNetbootin allows you to create bootable Live USB drives for a variety of Linux distributions from Windows or Linux, without requiring you to burn a CD. You can either let it download one of the many distributions supported out-of-the-box for you, or do it from your own iso file. I tend to use my own iso files, as some of the files unetbootin dowloads automatically seem out of date versions. But I have found out that about half the distributions iso I downloaded do not boot when you go the naive run unetbootin with iso and standard config route I outline below. I am certain many more would work with the custom settings, but I dont have the time or inclination to try to read up on each distribution’s bootloader modules to make it work. Thankfully slitaz isos have always worked via unetbootin.

* download from http://unetbootin.sourceforge.net/ (linux or windows). It will run straight from the downloaded file, no install needed.

* put your usb key in the USB drive and clean/format it (fat32 is fine)

* run unetbootin, choose the “iso” option and select the downloaded ISO. Wait till it’s done (can take a while if you have the “packages” iso)

* get to your mininote

Step 2: boot

* put your key in the machine and start the machine

* press F9 (a few times) to trigger the boot menu

* choose USB

* at the boot screen, normally all you need to do is press “enter” to start booting.

If you see the white/blue unetbootin boot list instead of the red/brown slitaz boot screen shown above, choose the “core” option to proceed

* it will take a little time as everything is decompressed and a memory disk is created

* you will need to set 3 choices: the resolution (slitaz does offer the native resolution of the screen, something which so many distributions failed at), the language and keyboard

* once you are on the login screen, just leave the suggested name (tux) as is, and press enter twice (no password)

* IMPORTANT mount the USB key first (else slitaz might mount it to an inconvenient place as you browse around and then you have to reboot and start over)

* at this point it is worth testing if the current packages for wireless work out of the box or if manual work is needed (...)

Step 3: install

The SliTaz install is rather simple - a few caveats though:

* you need to have your partition ready in advance (the install script will format, but it can’t partition)

* you need to have the USB mounted as /media/cdrom to fool the installer

* the grub bootloader instance that SliTaz installs does not include any existing partitions, just the SliTaz instance - if you have others, keep copies of your Grub menu.lst to copy the entries after the install is completed

Step 4: reboot into your new SliTaz

You will have to make the same 3 choices again (resolution, language, keyboard) and you get your newly installed SliTaz

If you have a wired connection, the first thing to do is update packages and get the wifi set of packages installed. This can be done from the command line or from the GUI package manager, I will show both ways in the next article. I will also list a few of the workarounds for some problems I have encountered at one time or another

I’m not a social media expert but with all the people going “no way corporates will pay for what they are currently getting for free”, I had to jump in with my contrary opinion. I think they would pay for the right package, and here’s why

1) Identity issues
the whole issue of identity management is important, and it *has* a cost and value - I suspect it can fly. After all, people pay for co-validated SSL certificates, they might pay for a corporate set of twitter accounts which required verification (and cost something) and would a)give people trust it is really the company it says it is b)make sure other people cannot pretend

2) Multi-user accounts
Allowing for 1 twitter stream shared by multiple people contributing, in an orderly manner, or to link multiple twitter accounts in some form of “twitter net”. And perhaps company private messages as well as public ones. Or groups.
Yes, some of those can be achieved with third party tools but corporates aren’t going to mess with that - they want a complete integrated solution.

3) Serious management tools
to archive and delete tweets, but more importantly to search and filter them. Yes, there are third party utilities, but most aim at a totally different public, the hobbyist or blogger, not the corporates who might have 30 minutes a day max for twitter itself at times.

4) better tracking.
I don’t mean the kind of keyword tracking that means you can detect and join conversations about you (that’s #7) , but ticket-style tracking, to make sure no incoming tweet gets missed (which is ok if you are a person but a PR nightmare if you are a corporation).

5) better integration
tools to integrate twitter with other company communication solutions, from their own websites and blogs to email and mobiles. Easily. With corporate languages and tools.

6) agents
i.e. ways to program twitter actions that trigger specific outside-of-twitter actions. Easily. With corporate languages and tools. Could be contests, polls, voting, orders, event booking, or meeting scheduling. Some of these can be done via third party improvised solutions, but again a corporation will want the whole package from one partner, with SLA and accountability. I think twitter sees this one

7) marketing support
- audience analysis
- audience segmentation

8) full business intelligence analysis
audience and message analysis
- user profiling
- provenance
- types of messages, independent, incoming, replies, retweets
etc.

9) filtering

I think 7) and 8) are likely to be outsourced to third parties for a while

First, as I have said here and elsewhere before, I love blipfoto.

It is a very simple but beautifully done photoblog, or photo diary. You can only post 1 image a day (but no pressure to do it every day). The community is friendly, encouraging, and creative. If you are looking for a way to encourage yourself to take more pictures more regularly, you really owe it to yourself to check blip out. It certainly did that for me, to incredible degrees at times. I lost it over December but fully expecting to go back.

The site has been free for 2 years. But like every site out there, they have costs, and need to start covering those costs.

They are not adding ads, they are adding premium memberships, with improved features and additional bonuses (I suspect these will be special discounts, perhaps also ways to show/export your blog but probably a lot of premium information, courses etc. But I am just guessing). But they are doing something else that is very much in the spirit of the community created around the site: they are creating special supporting founder memberships:

(from the blog
you’re going to have a never-to-be-repeated opportunity to show your appreciation for everything Blipfoto has done so far, give us a kick-start for 2009 and stake your place in history with an exclusive Blipfoto Founding Membership.

For £40, you’ll receive:

- 18 months full membership, starting when we introduce our membership option
- an exclusive founding member’s icon, which will stay with you forever
- a specially produced founding member’s enamel badge
- 10% discount on all future Blipfoto purchases, including membership fees

As if that wasn’t enough, when you take out your Founding Membership, you’ll have an opportunity to pay a little extra and lay your hands on a set of 200 personalised Founding Member Blipcards. Again, this is an exclusive, one-off print run which will never be available again.

Why do I think this is so clever? Note that apart from a few token trinkets, you have no idea what the membership features will be. They are not selling the premium features, they will have plenty of time to do this later. No, they are appealing to their core users, the ones who love the service and would pay for it as it is (even though the service as it is will always remain free). They are saying “if you like us, trust us, and fund the time we are spending creating the next level”. I think this is an extremely open and modern way to fund an application, and a very courageous approach.

It works for me.

Even though I have no income at the moment I will become a founding member of blipfoto, because I have received real value from the site and want to support its team to do more of the same. I am not doing it for the badge or pin, I am doing it to support the site. I want to see what it can become, and am willing to chip in to make sure it gets the chance to get there (note that this is the same reason I paid for lwa, 72photos, aviary and others, and I would do the same for a site like bookmooch or friendfeed).

Please don’t come and tell me that your particular OS is the bestest of them all, super stable, easy to manage, easy to learn, no security issues. It isn’t. None of them are. If you think so you have forgotten all the times you scratched your head or tore your hair trying to figure out how to do…

Originally posted on the donation coder forum, but I thought I’d also put it here :D

Note: this is not a rant against linux or open source, not by far! I was using slackware in 1996, that’s how far linux and I go. I am a firm supported of open source, but frankly it is not the universal solution and the best of everything just yet. It might be, but it isnt. There are warts. You make the case for open source with warts, not forgetting about them in near religious zeal. And you certainly don’t claim your OS is perfect, or claim that everything is linux…

It always bothers me when I see people get religious about an OS (or programming language) - this started as an open minded conversation and at some point it starts being an advocacy discussion - with people using the usual myths about each other’s OS (linux is not that user unfriendly and mishmashy, neither is windows that insecure or unstable). The worst is that most of the people get all religious not about the reality of their OS (or language) but the idea of the OS, and the image it projects about them.

Once someone gets religious, then others feel they have to defend their choice (even if they weren’t religious about it, their image has been attacked, implying that they are morons/heretics for using something else. Hard to shut up after that)

I have used: several flavors of DOS, Vax/VMS, SunOS, Solaris, Opensolaris, FreeBSD, NetBSD, OpenBSD, desktop distributions of BSD, Windows 3.11, Windows 2000, Windows XP, Vista, Windows Server (NT, 2003, 2008), AIX, OS/2, HPUX, SGI, MacOS, about 20+ flavors of linux over 15 years.

I have administrered/managed, in a commercial setting: DOS, SunOS, Solaris, FreeBSD, Windows (the whole list above except for server 2008 which i only played with), AIX, OS/2, MacOS, Linux: Slackware, Redhat, fedora, debian, DLD, Suse, Centos, and a few specialised one (router/firewall) i can’t remember right now. Some as servers, some as desktops.

So when people talk to me about how wonderful X is, or how innovative, I tend to see red.

First, If you cannot list at least 3 ways in which the other person’s OS is better than yours (things you wish your OS had) then you don’t know enough to debate in the first place. This is called Nebbe’s rule when applied to programming languages, i’ll call it iphi’s rule for OSes.

1. All OSes suck - they fall way short of what an OS should be and might be one day - but most of them don’t suck enough that we cannot get used to them and like them

2. All OSes are unstable - at least any one I have ever used with a GUI has had mysterious crashes, problems, freezes, when pushed a bit.

3. Updates and software install are a problem on all OSes. There will be numerous cases and people who have had things mess up just by trying to install or uninstall on any OS - whether windows, macos, solaris, bsd, aix, linux distributions. If you think you haven’t had any you either have been extremely lucky, or you have forgotten the teething problems in your enjoyment of the idea of your OS.

4. All OSes are insecure especially in the hands of an uninformed user. Granted, some are safer because an uninformed use cannot even begin to use them (this is not a positive feature in my eyes). All of them have vulnerabilities, some of which dont get fixed for a very long time.

5. All OSes are frustrating - With any OS, there’s a time right out of the box where they are fun. Then as you start to really do work with them, especially with deadlines, the cracks will appear and you will tear your hair out. Then if you stick with them you will get to the point where they are stable, work to your satisfaction, and you will be comfortable like and old couple. It can take a month or 18 to get there, depending on luck, the match between the chosen OS and the task you are trying etc.

6. All OSes are fun if you use them to dabble. If you use an OS mostly to have fun and dabble, without pressure, you will like it better. So if you used windows at work but linux at home, linux will feel infinitely more easy, fun, stable - because you can just put up or ignore things that are less than ideal, and what projects you conceive will be projects that fit within the limitations of your chosen platform. If you have linux at work but windows at home (for games and chatting), you might feel otherwise. I have at some point or another absolutely hated every single OS I have had to use, except for the ones I have only ever dabbled with.

Other things OSes don’t do right yet are: navigation and launching, I/O, filesystems, file management, multimedia (try having multiple sound OUT and wanting to send different things, or the same thing, to them!), encryption (all right, opensolaris’ zfs is getting close on this one), and the list goes on.

I’m probably not alone in having spent time sketching ideas out only to never do anything about it - not because research showed it wouldn’t be a success, but because I never got the courage or focus together to even get that far. In my case it has often been “don’t feel like doing it alone and cannot quite justify trying to lure one of my friends to waste their free time on it”. Others do just that, though, and succeed.

Some would not have worked, others certainly would have with the right partner, but I never even really looked. I need to get some of those to rest, so I stop feeling bad about not following up on them.

I’m sure many of you know that feeling.

The stealth PC silent pc shop

In 2002-2003 I considered starting something about silent PCs, because I hate background noise and couldnt find much in the UK. I bought domains, I played with platforms, I contacted manufacturers, I got a plan together. But since I was running a web agency at the time that took me 10 hours a day already, I couldn’t do it without harming that business, so I didn’t. Others did it, but now silent is mainstream and most of these stores don’t exist. I think there is still potential though, as many products are still hard to find.

This was revived in 2004 around tiny form factor PCs, again because they were so hard to find. Rinse, repeat.

Smartguilds.com

At some point I also considered writing an online guild management tool for online gaming guilds. Being a gamer I had ended up in guilds or playing communities more than once, and being the person I am I always ended up an officer/leader… So I thought there was a potential for a service offering a neat forum/roster/calendar/loot manager/wiki/chat server etc. Again I got domains (12 or so), sketched it all out, played with the platform. Then did nothing, I didn’t have a designer that I could call on at the time, and then I got a lucrative job offer. Others did it, but I can’t tell how well they are doing.

Cooklink.net

Another idea I worked on about 2-3 years ago was social networking around recipes - there wasn’t that much of that around, just the normal sites where you could submit recipes. We had the idea to create a much more agile, web2.0 way of entering recipes, and using simple semantic analysis to allow recipe exchange across languages - so for example Bruno in Berlin could enter a recipe in German and metric units, but then Sam in Colorado could see the recipe in english and with US measures and temperatures, and perhaps even some altitude notes. Played with a drag-and-drop recipe builder, and then with text analysis. Then I took a job with a media and semantic search company and left this idea behind. It would have needed a few smarter people around I think.

The multi-language and international connection is a hard thing to achieve in text based online communities (come to think of it it is hard to architect right even on a simple content site), but I think this would have been an interesting thing to pull off.  Content that is formulaic, for example recipes or craft patterns, or code is a great way to start, because it uses a limited, structured vocabulary and grammar which can be mapped across the language boundaries, allowing for some experiments in cross-boundary community building.

But the challenge is not just technical. Advertising and affiliate deals are usually limited to one language and territory. I have worked on sites in the UK where only 30% of the audience could be monetized (apart from automated network or adsense style ads). It takes a significant operation to deal with agencies in multiple countries, and I am always amazed how outdated this feels. Also, affiliate deals are again very narrow. Even amazon will only give you widgets for one country at a time. If you want something that shows the same books on a site but with the UK link to a British visitor and the German link to an austrian visitor, you have to build it yourself. And it is a pain.

I do hope this changes one day, I like a diverse world and being able to share without forcing everyone into one language. but this turning into a topic of its own.

news it ain’t

I am known for ranting quite often as to how much of the news we are given is not news. PR statements, discussion of what other media are doing, reports on studies, non events, curiosities, opinions - but not news. Not real information, and many real events left unsaid.

I thought it might be interesting to actually show how much white noise and fluff we are getting and rattle my cage every morning about what I saw and heard on the web instead of to my poor colleagues.

The world was spared.

Historyslices

That one I still haven’t given up already, although I cannot help thinking it is already out there and I haven’t found it, or could be built combining existing services in a large part. It would build on things like flickr notes and visual search and concepts from image recognition and genealogy sites and mapping - with a dollop of social networking to tie it all up.

This starts with your garden variety history buff. I know several, and I know most of those have accumulated boxes and boxes of images and articles. Some could almost build a museum. Most of them have a rather narrow niche of interest - the local train line, or their family, or a business, a town. They have done their research, and often can talk through stories, showing you old photographs and maps, connecting them together. And most have blanks - they have identified some of the buildings, vehicles, words on the images but not all. And for these blanks there might be another history buff perhaps 5 or 100 miles away, focusing on something different but who could recognize the unknown element, because it is familiar to them.

These people are not online much. They use ebay, know their way around a scanner, and know how to search for information online some. But they don’t publish. If they are very clued up they might be on a genealogy site or upload some scans to flickr.

What these people won’t do is spend hours and hours messing about with a fancy webapp. They won’t get the idea or time to put it all up on a site, even though they probably have enough to fill a site and a few books. But if there is a site where they can add things gradually and easily, which builds them a site and then connects with others (squidoo with semantics?),
upload just one image, register what you know, and what you don’t, in 3 or 4 minutes, then send it out there to see what it connects with (leveraging the commons, possibly?),

Because these stories will be lost if they aren’t moved into the digital age.

I really want this one, please tell me someone’s built it. Or, if you too have some history buffs in your family circle and have thought “there should be something online for them”, get in touch. I haven’t given up on or thought this one through yet. Since I first discussed that one Flickr has started the commons, and that is a step in the right direction, a building block. But there still is no connection or structured annotations or the ability for normal people to tie into it…

blumphster.com

A site to poke fun at all the ridiculous web2.0 ideas out there, which were surfing on the new bubble and hoping to print money. Was going to be a fake such site, but everytime I came up with what I thought was a silly web app idea that would work as a fake, I found a real one too similar.
Not particularly topical now that there is no bubble.

There’s actually a few more but I think that’s enough ghosts for today.

This is an update on my surprised report of problems with the ebay-paypal connection and how they hadn’t done a proper path for users when an error occurs, as mentioned in the previous article.

A week later they have created such a path.

The error is still occurring for me, but now I get a pretty page when the error occurs (it is still occurring and I still dont know what the problem is), and on this page a link that takes me to the old integration method between ebay and paypal. Exactly what they should have done in the first place.
Except it still doenst seem to work right because when I follow the link paypal considers me a new account, even after I log in, and wants me to enter card details. Thankfully I have found out that if i take 1 item at a time things work…

But to me it is fine - there will always be weird errors and things, and the thing is to not leave the user out in the cold, as I was with a text error pointing me to an unserviced email address…

Although I suspect they havent quite done it the right way, I suspect they have added exceptions to catch the error within the application and handle it within the coded app, but that for any unplanned error people might still be getting the basic text error. I hope not but human nature and development pressures being what they are…

Anyway to me the ideal way is to both
a) check the result of any transaction within the application and catch any error, and handle a recovery path within the application
and
b) not assume there won’t be fundamental errors that the code can’t catch (or that havent been thought of) which then push the user back to the server layer - and to therefore provide a path for the user on these standard error pages when that happens

Perhaps it is because I started 15 years ago when we were all webmasters and webmistresses and did everything from server configuration to menu graphics, but I keep being surprised of how many sites out there haven’t done the basic polish when it comes to simple server and service configuration.

Really basic, web hosting 101 details like having the site work both with and without “www” in front, or having custom error pages… these are still quite common even on high profile sites. Back in the early times it was caused by lack of skill, the industry was in a gold rush and many did not know how to configure things, some hosting companies were cowboys who often didn’t know much themselves, and the client would be told “can’t be done” or “will cost an extra 5000”.

I am not sure what the cause is nowadays, can’t be lack of knowledge. My guess is that it is a consequence of the increasing specialization - this something that falls between the cracks because it is not “development” or “design” or “content” or “marketing”, and nobody checks, or feels it is their remit to pick up the phone and try to get it sorted. So we still see horrible java dumps, or text based 404 messages, or are told that “blumphster.com is not accessible” (made up example, but i could name names!)

But even I was flabbergasted at what happened to me the other day… an example of how not getting the basics right has eroded my trust in one of the largest sites on the net… for not doing something that would take them 15 minutes per year!

Read More...

This is a set of commands and notes I used when installing BSD servers. I managed to recently lose it and had to go dig in old backups to recover it, so I figure I might as well save it here. It’s a list of steps and commands that can be used when on a new BSD5 machine - it contains port commands, compilation options etc.

This was based on a setup we had in 2005 which had light web servers (running small sites and perl ecommerce system), heftier web servers (at the time running php cms and zope/plone), 2 database servers (postgres), a mail server (postifx, popfile, amavis, dovecot etc.) and a separate logging server (gathering & merging the logs from all web servers,  other logs via syslog-ng and also running zabbix monitoring)

The version copied here is from 02/2006 - I have another one more up to date on the apache/php configuration option, but can’t be bothered to look for it at this time - since mostly I am keeping this for the crib sheet on port management commands it contains. And it seems easier to have it here, put on an older date but where I can find it anywhere. Perhaps it is of use to someone.

What is missing at the beginning is whole set of tricks to get the machine name, BSD version etc. from uname and other files into variable which could be used for the rest of the script. Because it was embarrassingly messy. So this script cannot be run as is, there are manual steps to be done in between.

Read More...