2014-05-31

HMRC Prompt Payment Discount VAT

Well, no sensible replies, even though they confirm the rules apply already for telcos...

So my latest email to them :-

OK, to make this simple I have a simple tangible question for HMRC on this new policy for prompt payment discounts.

I have, today, issued an invoice to a non VAT registered individual customer for telephony services.

The invoice does not mention VAT as a non VAT registered customer. It is for £14, but with a £2 prompt payment discount making it £12 if paid within a time frame specified on the invoice.

Previously the VAT would be on the discounted amount of £12 total, so £2 VAT. Simple.

With the new policy it depends on the "consideration actually received".

Please tell me what figure for VAT on this invoice you require me to include in the VAT payable on sales for my next VAT return?

I do not know if the prompt payment discount will be taken or not, and will not know by the time I have to do the VAT return (unless it is paid before then, obviously).

The rules you now have state that the VAT depends on the actual payment made. But I do not know what that will be. Will it be the discounted amount because paid promptly, or the un-discounted amount if not?

So - what do I put on my VAT return for the VAT element of this invoice? How much VAT should I pay HMRC in relation to this invoice?

Obviously it would be seriously negligent of HMRC to have introduced new VAT law and rules without knowing the answer to such questions before businesses have to implement them, and I am sure HMRC are not so irresponsible. I am sure you have an answer all ready for me and a VAT notice/leaflet and web page on the topic.

So, please tell me.

I look forward to your prompt, and formal, reply

2014-05-30

Long week (not keyboard related)

Well, more of an odd week to be honest.

We sacked that one customer for failing to read or listen, basically. No idea how that will pan out, but we have a missed appointment charge to apply and she will not be impressed with that one, I am sure. I may post a follow up. It really is rare to sack a customer.

We had a rather odd one with a new (prospective) customer that would not give his name. He just applied for service as "B". It was not that he is "know by the name: B" which I would have pretty much been impressed by, no, he has a normal name but refuses to provide it. He says banks are fine with Direct Debits under the name "B" (well, duh! they do not check the name). Sadly we are not. If ever we have to sue him, it will not work well trying to sue "B", even if we have his address. The strangest part is the massive rants he has over complaining about us and wanting to know the name of Lee's boss, and Alex's, and how he wants to register a complaint. He even hit companies house ton get my name and address. I almost wonder if I now have an official stalker? I give in! It is someone we are not taking on as a customer rather than sacking as a customer. Some times you have to say "no".

Then we have someone ordering new service on Office::1 but refusing to actually let us/BT install the new phone lines that are part of the service. Seriously, WTF? We're canceling that all and going to lose out on it big time.

Of course these cases scare the crap out of me because of ADR. Before ADR I would have no issue with telling any of them to piss off. I would not be the slightest bit concerned of a court case even. Now I risk an argument with an arbitrator and lots of hassle and costs. I hope none of them can go to ADR (the one with no name will be a challenge for him - no record of a customer with that name, sorry). But this is why ADR needs fixing. You can't run a business looking over your shoulder all the time to worry about total nutters.

At the end of the day it means a few extra words here and there on the web site and order forms just to be crystal clear. I am all in favour of improving communications and the services we offer, but I have to say that this week, or this month, has brought out the special people somehow - maybe it is the heat?

It also seems I have a new cousin, Logan. Actually no idea if a cousin - my sister-in-law's grandson, whatever that is. Welcome to the world, Logan - it is a strange place and will be a lot stranger by the time you can read this, I am sure.

I am loving my new kitchen, but Sandra made the mistake of ordering the cutlery tray... I am pretty sure we need Sheldon to install it - and I am unconvinced they have sent enough bits. [see pic]

Oh, and why the keyboard mention in the title? Well, having switched over to the FILCO keyboards rather than the IBM style ones, I am still unsure that this is a good choice, but it did strike me is that once again any blog post that mentions keyboards gets around at least 10 times the hits as any other blog posts, no matter how benign the post is. I may have to change the intro text in my blog to mention keyboards next and see what happens. It is bizarre.

We have has some progress on voice SIMs as we can now handle the character set coding and multi-part messages from SIMs much better. We are working on the other way too and have a meeting with the carrier in a few weeks. I am really keen to get roaming SIMs sorted next.

We had Zebra in as well this week. The developers over from the US, and a nice long chat. Impressed with the linux drivers and I think I managed to explain why you need them. I even managed to get the spec for their internal interface which I had previously only reverse engineered. They were pretty impressed with what I has managed without it! Nice to see a manufacturer taking an interest, so well done Zebra.

I have a school reunion later in the year - that should be fun - for the 50 year olds that went to PHS. I am printing the name badges for them.

Oh, and I got some new magnets to sort the door sensors. It is amazing what you can buy on line in magnets these days. You can't just order one - you have to get the 26kg N52 one to play with :-)

2014-05-27

Unifi UAP-AC with 802.3af and 802.3at PoE switches

This is just a warning to anyone using the Unifi UAP-AC Access Points and not using the supplied PoE injectors.

These are really good APs, and we recommend and sell them. However, I was caught out by a slight problem this last week.

We are running the latest beta test code for these. This is, of course, a risk we are taking. We fully understand that risk. But what happened was not actually an issue with the beta code.

We upgraded to 3.1.12 from 3.1.10 and encountered a problem. The APs kept crashing. We were lucky if they ran for an hour, and often way less. This made the WiFi almost unusable.

It had been fine on 3.1.10, or so it seemed, so we assumed it was an issue with the new beta s/w. Shit happens, we are running beta s/w, so not a surprise. Ubiquiti tried hard to find the issue. We sent logs and config and so on. It was clear from the beta forum that others had the same issue. The PRO models were fine, only the UAP-AC units were crashing.

We found the problem, to our surprise.

These come with there own in-line power over ethernet (PoE) power supplies. Some of the lower models are a passive PoE, but the PRO and AC units claim to be 802.3af and 802.3at compliant. To be honest we had not done a lot of PoE, and we got some nice PoE switches. What we had not spotted is that they are 802.3af only, and that is only 15W of power. The UAP-AC units need 22W so the 802.3at (25W) switches are needed.

Now, my understanding is that 802.3af and 802.3at should negotiate the power requirements. This should mean that a device needing more than 15W should not negotiate with an 803.2af switch and so should not work at all, perhaps indicating a power fault. Sadly, either the switch or the AP does not do this and it tries to work anyway. Somehow the UAP-ACs were working but the s/w update clearly increased the power usage and caused crashes.

Having worked this out, changing to use the provided in-line supplies fixes the issue. This is messy when we have PoE switches anyway, so we are going to try and find some 802.3at PoE switches.

It is a shame. If the negotiation logic had worked we would have realised our error when installing these, rather than much later on a s/w upgrade. As I say, I don't know if the switch (netgear) or the AP is to blame (though latest comments on the forum suggest it is the AP), but it is a cautionary tale none the less.

The complainers

I am intrigued by an advert for a new TV series "The complainers" about a trend of people that just complain for the hell of it.

I suspect I will be shouting at the TV.

At AAISP we really do try hard to provide good service for our customers (even though making this not part of the contract for legal reasons!). I know a lot of you know this, and appreciate it. But there are just occasionally, and very rarely, occasions where I really cannot win and we have to give up.

This is seriously rare, but I have once again (second time in 17 years, I think) sacked a customer. It is so rare, and such a concern, that I think it is worth blogging. What do you think? I am being a bit frank here, and obviously I am not going to name and shame the customer.

This is a case of a customer that is, apparently, an on-call doctor (surgeon, I think) and apparently needs to be contactable as a matter of life and death, but lives in the middle of nowhere that has no mobile coverage on any network.

OK, that is scary for a start. We really try not to be involved in "safety of life" services. We even make that one of our terms. We cannot exclude liability for death or personal injury from our contract terms, and if it happened and we were negligent we'd have to rely on our public liability insurance. The services we offer are subject to a lot of other companies (not least of which is BT) and is not something we can safely say will be 100% reliable. We'd love to say it is, but we are realistic here. We were not told this when the customer ordered (and would not have taken the order if we were).

So first issue, after the broadband was installed and working, was the phone line suddenly ceased. We suspect our customer migrated her number to a VoIP provider and that killed the phone line (which we were not providing) and so that killed the broadband (which we were). We are not entirely sure of the cause of the cease, but it was not us! We worked hard to get her back on line as soon as possible, obviously, and we started providing the phone line as well.

Thankfully, after a lot of nasty emails she agreed it was not actually our fault and we did everything right. It was hard work and took a lot of man hours to sort this, for which we make no extra charge.

We explained carefully that relying on a single connection when on-call was not really very responsible. We explained how we do offer some higher availability multiple line services using separate carriers and more than one phone line. It still cannot guarantee 100% uptime, but was likely to be a lot more reliable, especially in a rural area. But no, she did not take that.

This weekend she hit her usage limit on Home::1. Yes, for an on-call doctor using VoIP she purchased a domestic service with strict credit controls that cut off when you hit quota. We point this out specifically as a term when ordering, and on the invoice and order confirmation and in the printed information pack. We also explain that you can opt to have auto topup rather than being cut off. But no, she opted to stay with the "cut me off when I use too much" option, and it happened on a weekend. Fortunately someone was around at the weekend, picked up the email and topped her up and got her on-line, out of hours.

Apparently, doing exactly what we agreed is disgusting. She had considered going to a hotel so she could be contacted while the broadband was down. She wants compensation and will be taking legal action. Hmmm.

Bear in mind she has a phone line, and that works. It is a BT line we provide now (for the broadband), and has a phone number and can be called. If she plugs in a phone and someone calls the number, it will ring. We had explained this already after the last fiasco with the line cease. But apparently that is not good enough - she needs her working VoIP, and only that, on the one and only domestic broadband line she has which has strict credit control enabled.

I have said I will waive the six month minimum term and terminate the service on 30 days notice - clearly we cannot meet her requirements. I personally think that, if what she says is true, she has to be the most irresponsible doctor I encountered. Who goes on-call and then deliberately buys the most unreliable system of communications, ignores other means (phone line), ignores advice, and then complains? Does she go in to surgery with just one scalpal? I am quite shocked at this, to be honest.

I am struggling to find what we could have done better - I have checked, and we not only highlight the terms on the web page, and during the order, but we send email and paper information pack explaining it all in great detail. If there is any way I could do better I'd like to know and I'd like to improve our service. Sacking a customer is a sorry admission of defeat here, but I don't see an alternative.

Lets hope this does not turn in to an ADR case. I think we are being very reasonable letting her off the six month term, and wish her the best of luck with her next ISP.

I am a shark!

Thanks Mark for the unexpected post on ispreview, especially :-

"Suffice to say that the call reminded us a little bit of those scenes in Jaws ™ where the shark (Adrian) relentlessly picks off the crew of Quint’s vessel (Virgin), most of which are left to hopelessly flail about in an attempt to fight the inevitable truth of their demise; except in this version not even Brody survives."

2014-05-24

Virgin lies?

Virgin have stated to me (as a residential customer) that BT do not offer any "Fibre to the Premises" service, it is *all* Fibre to the Cabinet only. I pushed him on this point several times, and he even went away to check, but he is absolutely adamant.

Untrue[1] - BT do offer, in some areas, to residential customers. FTTP.

Virgin stated that they use twisted pair from the cabinet to the premises. That is what coax is, it is twisted pair!

Untrue[2] - coax in not twisted pair.

Virgin stated there is no-one I can take to alternative dispute resolution.

Untrue[3] - they use CISAS (as he said later in the call).

Oh what fun... So this seems to be a deadlock.

Recording is here (stereo)

Yes, I am being a bastard, but I am none the less quite shocked at the way they handled this. So, do I go to ADR? If I get compensation I'll donate it to charity to be honest - I am trying to see how the system works.

[1] http://www.productsandservices.bt.com/products/broadband/faster-internet
[2] http://en.wikipedia.org/wiki/Coaxial_cable
[3] http://www.virginmobile.com/vm/media/pdf/Complaint_code_Jan_11.pdf

2014-05-22

Not smartstamp, but what?

So, got a letter today, and this is the postmark/frank.

It is not a normal old fashioned franking machine as it contains a barcode.

It has Royal Mail logo but not the usual second class mark.

It has an IEC16022 datamatrix barcode like a smartstamp, but a different geometry variant. The content are a very similar style to a smartstamp, and include date and price and so on.

I assume it is some sort of franking system - but what?

Anyone know?

Update: Wow, after all of the hassle we had trying to get a spec for smartstamps, it seems MailMark is a system that RoyalMail allow systems integrators to work with to make mailing solutions. They have a process for becoming a producer. I think we may have to look in to it - to replace our use of smartstamp and make a sellable system!

Is RIPA fit for purpose?

As some of you know, The Regulation of Investigatory Powers Act 2000 provides a formalised process for authorities (e.g. police) to request information from the likes of telcos and ISPs. It also covers a load of other stuff including interception of communications.

Note: AAISP do not have any equipment connected to our network or in our network to intercept communications under RIPA or other such legislations. I.e. we have no "black boxes". Feel free to ask me on irc if you want to double check.

This process involves the police sending a form, usually by email, to the telco/ISP. The form does have details of a police officer that we can contact to check the form is valid. The whole process is, as I understand it, confidential, and could relate to investigations in progress, so naturally I cannot go in to any details on any RIPA requests we have had.

Being a relatively small ISP and telco we get very few of these. Maybe a couple a year. But we recently had the opportunity to see how the process works from the other side, i.e. as the victim of a crime.

As an ISP we have always suspected the whole thing is a mess of bureaucracy and delay. Most RIPA requests we get are not sensible (and as I say, we get very few). Some are plain wrong, e.g. one recently where it says "Communications provider: Talk Talk" and indeed was asking about a TalkTalk retail IP address, but was emailed to us not TalkTalk!

The requests we have had are either about a phone number or an IP address. I am not sure we have seen one for an IPv6 address yet. For a phone number, we could potentially have billing records for calls to/from the number, but we try not to hold any more than we need for billing and diagnostics, and we have not been required to hold data under The Data Retention (EC Directive) Regulations 2009. So, in general, we rarely have more than just billing name and address details. For an IP, it is much the same as IPs are fixed to one customer and we don't log what web sites people visit, etc. In most cases this is all the police need anyway.

Obviously we always stress that the billing contact may not be the user at the time, and that the installation address may not be the address the number or IP was used. We allow L2TP login for all DSL line IPs from anywhere in the world, and people can (and do) run relays, VPNs and TOR nodes.

When it comes to phone numbers we usually find the number is not in use, and often never been allocated. Spoofed CLI is very common in crime, it seems, and the police have a really hard time understanding that you cannot trust a CLI.

Now, when it comes to the other side, just after we were robbed (next day I think), one of the stolen machines did a bit of a phone home, logging in to dropbox. This was a staff member's windows machine. The apple boxes that are supposed to have tracking and so on, not a peep. Sad. This meant we know a Virgin IP address and told the police right away.

From what we can tell this was an address in Slough where someone "fixes" the machines - presumably wiping and re-installing and so on. An essential process in the resale of stolen goods, I am sure. If the police had gone there right away they may have found the stolen machines there.

Unfortunately it was a much much slower process. The police officer handling it had to talk to another department about tracing the IP. It was a process that involved some days before the came back asking the time zone. We said UTC. Many more days later they came back with "what do you mean UTC? Is that a time zone?". It was shocking. The time stamp was only to the minute, which caused an issue, even though Virgin IP addresses are sticky enough for that to be one address only. It was very frustrating.

I ended up contacting Virgin via my contacts saying "Please can you help this police officer fill out a RIPA form that you can process?". I don't know if that helped or not.

It was weeks before the process finally gave up an address, and an arrest was made, but the kit had all gone.

Whilst I am massively in favour of due process, I am not in favour of broken bureaucracy. I don't know why there is not an internal police web portal where the investigating office completes details, maybe it is flagged immediately to a superior officer to approve, and then sent electronically to the ISP, with an electronic reply. Large ISPs could even have some digitally signed XML interface to handle the RIPA requests and reply in seconds, but will all the approval process, authentication and paper trail that is needed. If that had existed we would have probably got our stolen stuff back and they may even have been able to catch the actual thieves when they tried to collect the stuff.

Oh well.

Speechless

I am rarely speechless, but I honestly do not know the best way to reply to an email I just got.

One of the people we were taking to court for spam, and got as far as a hearing date even, and whom I posted various replies and comments on my blog at the time, just emailed me.

Naturally, seeing who it was from, I was concerned. I never expected to hear from him ever again.

He is after buying from us, or us recommending, spam filtering services to sell to his clients.

I really don't know what to say.

Maybe he has turned to the light side?

2014-05-21

Non-conformist

As I am sure some of you know, I am a tad non-conformist. Not in the same way as all those other non-conformists, obviously.

Some times this can cause a slight nuisance.

For example, I have a short email address, only 6 characters in total. It is, of course, fully compliant with the specification for email addresses (the relevant RFCs) so not actually non-conformist at all. It is, of course, unique to me, as are all email addresses, so again not really non-conformist. There are many others who have such a short email address (everyone in my family for a start).

Sadly, lots of web sites make assumptions about what is a valid email address and will not accept one so short. This is just sloppy. There is a standard for email, and it is not that hard to validate against that standard. There are tools to do that. But instead, designers assume they know what an email address looks like and make up their own broken rules. Getting this fixed is hard work, people simply do not care. (In some ways I can understand - even I have a number of minor niggles in our systems still on the list to get corrected, but we do try).

Fortunately, with email, it is relatively easy to have an alternative email address. I have some in the domain stupidwebsitethatdoesnotunderstandshortdomains.e.gg and rfc2822.me.uk and temporaryemailaddress.co.uk that I use some times.

Many people have the same problem where web sites refuse to accept email addresses using punctuation which is specifically listed as valid in an email address in the RFC. Some sites will not even allow a simple plus character.

The other problem I have is my mobile phone number. It is a Bracknell 01344 number. I can make and receive calls on that number using my iPhone, as well as sending and receiving texts. It works in every way like a mobile. Whilst this is a tad non-conformist it is not new. When Orange launched over 20 years ago they offered London and Birmingham landline numbers as an option (for an incoming call charge). I very nearly got one of these at the time, and wish I had.

But, once again, people assume they know best and expect a UK mobile to start 07. One culprit is Barclays - I am unable to use their mobile banking app or pingit (and hence paym). This is annoying, and unlike the email example, I don't have an 07 number I can use instead, not any more.

But, for both of these issues, we have a cunning plan. I am pretty sure I am not the only one trying this at present as it was discussed on irc I think. The plan is simple. Sign up for a service using an email/number that they will accept (which is where temporaryemailaddress.co.uk comes in). Then require them to correct and update the information they hold on me on their computer systems in accordance with principle 4 of The Data Protection Act 1998:

 "Personal data shall be accurate and, where necessary, kept up to date."

The data controller has an obligation to adhere to the principles as per section 4(4). So, basically, it is the law that they correct the email/phone number that they hold!

I am just writing to the data controller at Barclays registered office, having got a Tesco mobile SIM (which comes with 10p credit on it) and using that to get mobile banking and pingit on my iPhone before throwing the SIM away. Now I am asking that they correct and update the mobile number to my 01344 number. We'll see what happens.

In the mean time, I can report one success. The DVLA have web pages (I think for tax discs and the like) and will email a confirmation to you. They did not accept the short email address used by my father. As this was only for the confirmation email and not data they had stored, the DPA trick would not work, but it seems my father is good at pestering, and apparently the DVLA do now accept his email address - well done Dad!

Also, I was impressed with Uber as I had no issue signing up and getting the confirmation text on my 01344 mobile number. So some people can do it right.

2014-05-20

Slower fault repair?

As reported, OFCOM are working on new rules for Openreach to fix faults within 2 working days.

At present, for broadband, the target fix times are 40 clock hours, or 20 clock hours for a higher priority maintenance, or 7 clock hours for the highest maintenance option.

40 clock hours is always less than 2 days (48 hours), and 2 working days could be 120 hours over a bank holiday weekend.

I really hope BT don't use OFCOMs new rules to change to longer SLAs than they currently have in place.

(Note, BT rarely actually work at the 40 clock hour target, especially over weekends, but if they had a target of 2 working days, it would clearly be worse).

So really, what we need, is not "quicker fault repair" but more of "actually work to the targets already agreed". What we need is much more substantial compensation for failing to meet the targets they set themselves in the contract.

We need BT to actually work to current targets and pay if they don't!

At present, if BT miss a 40 hour target they pay a tiny penalty to the ISP, literally something like £1 a day for 5 days. Then, they pay no more - so even when 40 hours is missed there is no urgency at all. They should be paying tens of pounds an hour for missing a target, in my opinion, as that would maybe motivate them to work to existing targets. We have this crazy situation where, if they miss an appointment, they have no system to not consider the fault extra urgent after they have screwed up. They don't even deploy the same extra effort as they would for "7 hour fix" fault once a 40 hour fault gets to 33 hours in, which seems broken to me as that is what the fault is at that point!

As an aside, and we have asked BT to explain this, BT have admitted that the 7 hour fix option from BT Wholesale apparently does not have any SLA for the engineers working for Openreach to actually try and meet that 7 hours! This is a somewhat daft thing to say as the engineers working for British Telecommunications plc trading as Openreach work for a company that has a formal contract with Andrews & Arnold Ltd to try and fix within 7 hours. So there is an SLA in place on those engineers. BT try to pretend that is not the case, but they cannot get away from the fact that they are one legal entity and that legal entity has a contract to which they are expected to adhere. We await their response on that one.

Your hold music is too loud!

We received a complaint that our hold music is way too loud.

Apparently a deafening blast of Blondie, "hanging on the telephone". The poor chap jumped because it was too loud, and managed to hit is head somehow.

Obviously not happy.

But there is a slight catch.

We don't have any hold music!

What we have at the office is a quiet comforting "beep beep" every 3 seconds. Certainly no Blondie!

But the phone system passes the "on hold" signal through SIP. It even finds its way out through the PSTN, ISDN and even mobile systems in some cases. So you get "on hold" on the screen of your phone even. This is not always the case, some types of interconnect lose the signal, but in many cases this gets all the way to the far end phone system.

What this means is that the music on hold this person heard was his own phone system's hold music.

Ooops.

2014-05-19

One SIP at a time

Some more technical challenges with SIP for our new SIP2SIM service...

We are acting like a SIP client, i.e. like a SIP phone, registering with a server or VoIP provider. This could a system in an office, e.g. asterisk based, or even using one of our FireBricks as an office PABX, but it could also be a commercial SIP/VoIP service that is designed to work with the likes of a snom phone.

To do this we have to not only comply with SIP specifications, but all of the quirks and security restrictions that VoIP providers have dreamt up.

We also have multiple call servers our end, and we want to keep it that way to allow us to scale up to more call servers. So we deliberately started with two servers (and a third test server) so we solve the problems we find as we go.

Balancing the load.

We split the SIMs that are registering between the call servers using a hash, and based on the currently in-user call servers. This keeps the registering server pretty consistent. When we change call servers we de-register SIMs and re-register from a different call server.

Asterisk losing registrations.

We seem to have a snag with asterisk that will see a new register and a de-register from previous call server, and may see them in that order and decide the client is no longer registered, even though the de-register does not match the previous register or even come from the same IP. This is asterisk being an arse, sorry, and we have yet to work out a sensible fix for this. It works on next registration, and obviously calls from the SIM work. We do maintenance rarely and will probably move to overnight changes to minimise the impact of this.

Asterisk boxes deciding to challenge.

We hit a snag with asterisk boxes, as they tell if they need to ask for credentials (challenge the registration or invite) based solely on IP. Well, the config takes a hostname and they look up just one IP for that hostname and check that. This means that we cannot send calls to a customer from one IP address for both authenticated (i.e. we expect to have the call/registration challenged) and non authenticated, such as calls we have for their service from the PSTN where they may simply trust our IPs and not take a user/pass.

To fix this we gave each server two IPs, one for authenticated messages and one for not (also two IPv6 addresses, obviously). This allows a distinct config for both types of INVITE message on asterisk. Messy, but works.

Register and call from different IP.

Incoming calls to the registered contact, and outgoing calls from a SIP, should ideally be quite independent things. There is no reason they need to come from the same IP address. They both have credentials that are challenged to confirm they are genuine.

However, it seems some VoIP services will not accept calls unless a registration has been made first, and then only accept calls from the same IP from which the REGISTRATION is made.

This means we now have to bounce calls to the call server that does the registration for the SIM in question before sending to the VoIP provider. Annoying, but works.

Providers editing the contact URL!

We send a contact that has a sip:user@ip part and an attribute ;line=xxxx as part of the contact URI. This is a simple and valid technique used by lots of devices. Some send ;instance=, and some send all sorts or extra tags in the URI. It is a way for the client to be sure that calls they receive are genuine (even if from a different IP to which the registration went).

Sadly, at least one provider strips this from the URI. Why? This is not a valid thing to do!!!

OK, so we have a work around, we now only register a contact as sip:user@host, where the host has a prefix and we use wildcard DNS to ensure that is the IP or our call server.

We did run in to a snag with one server expecting INVITEs from the IP that they resolve the contact host part to, and we were setting the wildcard URL to the main IP of our server not the IP that we use for authenticated requests. We have fixed that now, and we think we finally have a reliable way to register to a wide range of VoIP servers now.

Well, until the next snag!

2014-05-17

How did Windows get to advertise on my keyboard

Do Microsoft pay the keyboard manufacturers to advertise.

To be fair, I know these filco keyboards can come with blank key tops and replacement key tops, but I thought I'd just design and print a blank key to replace the windows key anyway.

http://www.thingiverse.com/thing:334424

2014-05-16

Lies, damn lies, and statistics (internetmatters.org)

I am all in favour of more education, and evidence based policy and so on, but when you get an officially endorsed web site like www.internetmatters.org misusing stats like this you have to wonder if there is a hidden agenda.

This stat could mean that 100% of 18 and 19 year olds have seen on-line porn and zero below 18 have seen any. It could mean anything. It depends on the age distribution of the sample, and all sorts. What is worse it is seen porn online, that could be anything from having seen page 3 of the sun on line over the shoulder of an adult at a PC, or some unwanted popup advert because they have not set sensible browser settings. It could mean anything.

As a stat it is completely meaningless. The age of people does matter - if there are 9 year olds actively watching online porn (or any porn) I would be very concerned. If we are talking of a 15 year old boy I would be totally unsurprised. If we are talking an 18 or 19 year old ADULT watching online porn then what is the problem at all?

Quoting that stat at the start of the video ruins any and all credibility of the rest of the video. What are they thinking?

Update: I was thinking more on this awful stat, and a couple more points come up, purely intuitively. 1. The taboo nature of the subject mean that even adults will deny seeing porn even when they have, but it is very unlikely someone would say they have when they have not. 2. The combination of both being taboo, and seen as wrong doing for a minor would mean children are far more likely to deny seeing porn than admit it. This means that the 57% who admit seeing online porn is probably a gross underestimate. 3. Even so, if 57% have done this terrible thing and society is not falling apart, that is a serious indicator that there is no actual problem to solve here. If the stat was 99% or 100% would we really not have to say "clearly this is perfectly normal and nothing to worry about"?

Thanks Alex for pointing it out to me.

2014-05-15

New Kitchen

We got a new Kitchen! Here's the story...

The excuse was that I got a new cooker, and the colour did not match, but to be fair the old kitchen was around 14 years old, and some doors had come off cabinets, and it was not that good a design anyway, and well, you know how it is. Why we had yellow walls I do not know. I am a sucker for doing whatever Sandra wants.

We went to Wickes. They have run a lot of adverts, and are convenient and local and have a showroom to see how all the different styles work. They came and measured and then did a design on the computer with (horribly slow) rendering of what it would look like. Sandra sorted all of the details, of course, though oddly she ended up with things like "I am sure he said this unit here would be a pull out thing not just shelves" so maybe some misunderstanding in the process. I paid for the units and everything there, and the installation is paid separately at the end.

It seems Wickes subcontract to a local installer, a chap called Mark from Wokingham. At this point I would recommend him, his number is 0118 977 2405. He is the installer, but he arranges all of the other tradesmen needed. Finding a good tradesman is a nightmare, and we have found Mark and he knows good local tradesmen. They were all good, and we needed a sparky; a builder; a plumber and gas man; tilers; and decorator. They all did a good job and Mark sorted it all with one bill at the end. The bill was very close to the original quote with changes as we had changed what we wanted.

The only bit not part of the Wickes order was the lights. I did that. Of course anything I do would be wrong, being married I know this, but what was handy is that when I put them up and they did not look that good, Mark suggested a different way to fit them, and as it was not my suggestion it was right. Thanks for that Mark. I drilled the panels, and got some longer screws for the plasterboard fixings. I had to 3D print the white screw caps as I could not find any. I had to drill a hole in the ceiling for the transformer. The end result is great, and it was not actually that hard to install them - they save power at only 45W each but are very bright. They give the impression of skylights.

For the floor, Sandra went for patterned tiles in a somewhat Moroccan style. Bold choice, but I think it worked well. The tile supply company also recommended the grout to use, some flexible grout the tilers had not heard of, and sadly it shrunk and cracked the next day. They ended up scraping it out and re-pointing the floor. Moral of story is let tilers pick the grout they know.

For the walls Sandra originally wanted all of the walls tiled wall to ceiling. That was bold indeed, and expensive, but Mark persuaded her to go for only half of the room tiled, but she did go wall to ceiling, which worked well. The tiles she chose were a tumbled marble mosaic tiles, i.e. 6x6 of them on a sheet. Apart from not being cheap in the first place, if ever you go for these you have to consider a few gotchas and tell the tiler before getting a quote. For a start, being a natural stone, they were not sealed. This mean they had to be washed, dried, sealed, dried, and sealed again and dried. This adds time (and cost) to the job. Also, they do not go up like normal tiles - you cannot score and snap them, you have to stone-cut the tiles and fit like a jigsaw puzzle. This adds more time and cost. Then, finally, as they are separate, you are looking at around 6 times the normal amount of grout and time to do the grouting. Using mosaic natural stone tiles probable doubles or triples the tiling job and adds a lot to the cost - but they look bloody good. There is another catch - do not let them get wet before they go up as they lose adhesion to the backing material and a sheet of tiles becomes 36 separate small tiles on the floor when you pick it up. Thankfully only the bottom sheet of some of the boxes got wet and they needed the individual tiles to cut for edges and around sockets.

The work tops are solid oak, which means they will need oiling every month or so. No idea if that will happen, but they are pretty impressive none the less.

Mark was able to use some of the worktop to make up some additional shelves we wanted that were open and not really a traditional kitchen unit.

As part of the work we realised that it would make a lot of sense to move a door from the end to the side of the kitchen. Hard to explain but the result is much better. The new door way turned out to be in a supporting wall and not just wood and plaster, so took a while to knock out, and needed a lintel.

The result is we now have a door to the kitchen off the hall and easily accessibility for the front door. I am very pleased with the change, but obviously it added a few hundred to the cost.

This did mean the sparky re-doing light switches and adding a new light fitting in the new corridor to the utility room.

It also meant creating a new door step out of worktop wood.

The units in the kitchen are standard Wickes units, which are chipboard with laminate. This is standard stuff, but I was quite impressed with the build quality. They used full chipboard sheets as backs and not a thin sheet as I had seen before. They also have mechanisms for the drawers that rival what we see in server racks for the rails. The corner unit is a feat of engineering pulling the hidden corner of shelves out when you open the cupboard. All doors and drawers have a soft close thing that looks quite cool.

Sandra also decided on a new radiator. She wanted something that would allow her to dry tea-towels, and the like, and the old radiator was not really good for that. What she went for was a rather old style and I think looks quite cool - or hot - or whatever.

They used shiny copper fittings, which look OK, but I expect in time we'll end up painting them white when the get tarnished.

As you can see, we also went for an old style butler sink, which is nice. The tap pulls out on a hose so you can spray things, but seems to be much lower pressure for some reason.


The end result is impressive, and we now have the fridge and freezer and dishwasher and microwave and kettle and toaster all back in the kitchen and Sandra is working out what will live where.


Please Barclays: Email machine processable daily transactions

So, Barclays have an ideas site.

It has a way to submit ideas and for others to comment and vote.

Please can everyone that thinks this is a good idea visit the site and vote.
I don't think you even have to be a Barclays customer.

https://www.yourbank.barclays.co.uk/t5/Your-everyday-banking/Email-machine-processable-daily-transactions/idi-p/19947

We need to load banking transactions every day and it is tedious to get them from on-line banking. Why not have a simple option to email an encrypted copy every banking day for the day's transactions. I really don't mind whether you do proper PGP encryption or just a password protected zip file, whatever you like as long as it is machine processable. This is easy for you to do, eliminates any load issues as is a batch job and outgoing through your firewalls 

Thank you.

2014-05-14

VAT change on prompt payments

OK, sorry, this is a bit technical and accountanty, which is the worst sort of anything.

There is a concept of a "prompt payment discount" that you can put on an invoice for goods or services. In principle it is simple, pay within a certain time frame and get a discount.

Now, this has all sorts of problems. We used to do this on broadband back in the day, and people got confused. It is simple as a concept, but some people simply could not cope.

Apparently these are really common in construction, and what was fun for us, when charging such people for broadband and the like, is they routinely would both pay late and take the discount. This did not work with us, and we considered the late paid invoices under paid, and so the next invoice was even more underpaid and/or late.

What is fun on these is the VAT. The concept is that the extra you collect when late is a financial charge and outside the scope of VAT just like late payment penalties, so regardless of whether it was normal for people to pay on time or late the VAT was worked out on the discounted amount only.

Even so, the number of people confused by the idea of paying late meant paying more led to us changing our systems to actually issue a "lost prompt payment discount" invoice (with no VAT) once the payment was late. It was the only way to make them understand. The issue was that the original invoice changed. It was for one amount when they got it, but by the time they paid it, it was for another amount, and they had no way to change their accounting system to allow for that change. The separate charges for lost discount made it easier.

Anyway, in the end we stopped doing it - we do statutory late payment penalties for commercial customers and interest for consumers that pay late. They are both separate invoices issued later. Simples.

Where it gets odd is some new VAT changes. We got a VAT briefing on this recently, and I had not seen anything on it before. For the most part these changes come in next April, but apparently for some industries (telcos) it came in in April this year!

The change is all to do with aligning with the EU, apparently, but I don't get it. It also says it is prompted by the number of telcos doing this for consumers that are not VAT registered. HMRC clearly feel that they are losing out.

The change is that the VAT relates to the actual consideration received. This is a massive change as it means the VAT on an invoice issued is no longer set in stone. If I have an invoice with a prompt payment discount the VAT will depend if paid on time or late.

How the fuck does one allow for that on a VAT return? The VAT due depends on what you have invoiced, but you have Schrödinger's invoices here - the VAT is not known until the payment (or at least the deadline for payment) is reached, some time in the future.

I have emailed HMRC to ask how the hell we do this, and will be interested in the reply.

What makes this all so pointless is that, as far as I know, late payment penalties are still non VAT, so anyone doing this simply changes to lower invoices (and VAT) with late payment charges instead of prompt payment discounts. Everyone pays the same but HMRC do not get more VAT. So why the hell change the rules?

Update: HMRC actually said:-

The measure has been introduced for supplies of telecommunication and broadcasting services where there is no obligation to provide a VAT invoice from 1 May 2014.

As there is no obligation to provide a VAT invoice - i.e. supplies to people not registered for VAT - it should not be necessary to alter any invoicing at this time.

Which is mind boggling. Just because the invoice itself is not a VAT invoice does not change the fact that the supplier has to account for the VAT correctly and complete it correctly on a VAT return, which could be for end of this month - yet HMRC had no clue how they expect the system to work and are only just starting a consultation.

Lies, damn lies, and statistics (British Gas)

I did statistics at school (yes, proper stats O level!).

But when I see crap like this I do not really know what to say.

One thing you really should avoid doing is using a false origin when showing something graphcially.

There are circumstances where the absolute level is unimportant and the relative scale over many things can be shown on a false origin, maybe, but not this case.

This is a simple "how does your gas use compare to last year", and shows last year and this year in some graphical way.

I looked and thought:

  "shit, we are using twice as much gas this year, WTF!".

Then I realised the coloured bars were meaningless. In fact we have used 0.84% more gas. Drawn on a true origin, that would have been obvious, and no surprise.

Why even have the coloured bars?

The electric summary is just as bad - looks like we have halved our electric usage when in fact it is down only 12% (though that is not bad).

Sadly both of these are based on estimated readings, so pretty worthless, and rather pointless sending me the summary (in two separate envelopes).

2014-05-09

Do CISAS know the meaning of "dispute"

A while ago I was involved in a useful meeting at ISPA (Internet Service Providers Association) along with a number other ISPs and OFCOM and CISAS.

CISAS is one of the two ADR (Alternative Dispute Resolution) providers of which a telco or ISP must be a member.

The meeting was useful. CISAS did clarify their processes a lot, and confirmed that there was an initial stage of confirming a dispute was "out of scope" at no cost to the ISP. They also have an option to settle at that stage for a lower fee than the full dispute fee. Both are useful, and I appreciate them taking the time to clarify the process.

We discussed many points, and made constructive suggestions to help the process and make things clearer for ISPs as well as complainers.

However, I did raise one issue which was the possibility of a matter taken to ADR which was out of scope because it was "not a dispute".

This created some confusion - as they say that they will take on any complaint, defined as an "expression of dissatisfaction". However, I pointed out that someone can be dissatisfied and have a complaint but it not be a "dispute".

The scenario is simple - a key part of the meaning of the word dispute is that it is where two parties disagree. This is part of the dictionary definition. But someone can complain where there is no disagreement. If someone was to say "you provided very bad customer service", and we say "sorry, yes, we did, we appreciate your feedback and do not in any way dispute your assertion that we provided bad customer service, we'll try and do better in future", then in such a scenario a complaint exists but not a dispute. We do not dispute what the customer is saying, so no dispute exists or needs resolving.

[The separate issue of saying "I want money for bad customer service" and us saying "the contract is clear that no compensation is due for bad customer service" is a contract dispute and not something CISAS can take on]

I argued that a non dispute complaint is out of scope as they are there to resolve disputes. The clue is in the name ADR (Alternative Dispute Resolution), and I think I actually used that phrase.

They really struggled with the concept, and I asked if they know the meaning of "dispute". They tried wriggling saying that they understand what I mean by dispute, and I pushed them with "do they know the dictionary definition of dispute?". I was somewhat assertive, shall we say, almost rude, sorry.

Basically the answer is no, they do not. They could not define dispute in the meeting. CISAS cannot tell me the meaning of "dispute" even though it is part of the term that defines what they do "Alternative Dispute Resolution".

I have given them ample time to come back to me with an answer to this and not seen one yet, hence now blogging. Sorry, you had your chance guys.

If an Alternative Dispute Resolution provider cannot even define the meaning of "Dispute" then there is clearly a problem.

We'll see if they have a reply now I have blogged.

2014-05-08

SIP2SIM

We know how powerful SIP2SIM is, and as well as working with home and business VoIP servers we are keen to make sure it works with public VoIP providers.

Today we sorted a slight issue with calls to a VoIP provider. We have a test system in place that allowed us to debug the problem for the customer and deploy a fix rapidly on to the live platform once confirmed and tested. It related to timing of sending an ACK after a 407 proxy auth error response. It has taken a bit of work to track down, but has been fixed.

What is slightly amusing is that this is a VoIP provider in Israel. This means someone walking around the UK has an Israeli number as their mobile phone number. Brilliant!

No problem with this at all - feel free to connect your SIP2SIM to any VoIP provider, and if you have any issues, talk to us and we'll work to resolve them.

It just amuses and amazes me what one can do these days, even when it is us doing it:-)

2014-05-06

Zebra ZXP8

I am pleased to say that Zebra have finally sorted my rogue printer once and for all by completely replacing it with a brand new one. Thank you, Zebra. You can see previous blog posts for details of the issues.

This is good news as once again we have a smooth operation for printing and dispatching SIM cards. The printer has a contact station which allows us to talk to the card as part of the printing process. In practice, all we need is to read the ICCID, but we could do whatever we want.

This means the printer is a sensible way to process cards for dispatch, even if we are not printing them - i.e. reading in from a hopper of cards, talking to the card, and sending out to the output hopper. We have the option of diverting the card to a reject bin if the card won't talk or we are not happy with it (including the possibility that someone put the wrong type of cards in the hopper). Obviously we normally combine this with printing on the cards.

There are a lot of card printers on the market, typically aimed at printing ID badges and the like. The Zebra ZXP8 is a tad special though. It is one of the only printers we can use on SIM cards.

Apart from the contact station, which is obviously important for the SIMs, the important feature is that the printer does not try and print directly on to the card.

Most printers do print directly to the card using a thermal transfer process. This process is very effective and allows 300dpi full colour (256 levels of Cyan, Magenta and Yellow) which gives a photo quality output. They usually also do black as a separate layer using some sort of resin that is thermally transferred (1 bit per pixel), giving a crisp sharp and very black extra layer that is ideal for bar codes and text. There are, however, some problems with printing directly. It is not easy (for some reason I do not understand) to print edge to edge using direct thermal transfer. It also works well only on a good smooth flat card surface. Obviously blank ID cards are a smooth surface like this and ideal for direct printing.

SIM cards are different. For a start, the card itself is not the same - it is not a smooth surface. It usually has some texture to it. It also has cut-outs around the SIM, and of course the SIM contacts themselves as well as an uneven surface on the rear of the card behind the chip. What we found using an Evolis printer was that we could just about get the black ribbon to print on the cards. This depended on the type of card, and a slightly different material in the next batch meant the black ribbon came out a feint grey. Colour was almost impossible. Hence we got the Zebra.

The Zebra ZXP8 has two important features to work with SIMs. The main one is that the printing, which uses the same thermal transfer and colour+black printing, does not print on the card. Instead it prints on to a transfer ribbon. The card is then sandwiched in the transfer ribbon between heated rollers applying pressure. The end result is that the transfer ribbon is fixed to the card leaving a smooth glossy surface and the full printing colour (under the glossy surface). This looks really good, and because the transfer ribbon is larger than the face of the card (slightly) it allows printing properly edge to edge.

The other feature is that the ribbon can have an inhibit panel in addition to colour and black. This sticks to the transfer ribbon and pulls it off the backing selectively. When the transfer is applied to the card this leaves gaps. These are ideal for ensuring no print or glossy transfer ribbon is applied to the SIM contacts, or a mag stripe or signature strip. In our case we clear the SIM contacts and the cut-out. If we print over the cut-out it creates a sort of cellophane effect over the gaps that comes off and sticks to your fingers. Similarly, if we print over the contacts or a mag stripe then it does wipe off. But that is messy, and the inhibit works really well to make a professional looking card.

The final part of the puzzle was drivers. I ended up writing my own as they only had drivers and a development kit for windows. My drivers are a simple linux command line which I have published for free. This allows image files or postscript for front and back for colour, black, inhibit and even UV layers (depending on the ribbon) and sending to the printer. It also allows a real time update of print progress and display as javascript commands to updated a div in a page so can be used from a web pages nicely. It also has a preview mode to allow the layers to be merged in to a final image instead of printing. We use that to make all of the card previews on our ordering web site even when customers are uploading their own artwork. The drivers even allow me to update the display on the printer to advise progress, and issues like wrong type of SIM inserted, etc.

I'd like to thank Jim in their Bourne End office for all his help and patience on this, and I can definitely recommend these printers to anyone doing SIM cards. Do read up on the meaning of "contact station" though.

2014-05-04

Bank Holidays

So, this weekend, I have done fuck all, so far.

I have sat around watching series one of Stargate Atlantis yet again, for want of something better for do. I have finally come in to the study and done some coding as doing nothing is a tad boring.

Why do we have these bank holidays? It is not like banking stops!

I should go code some EAP stuff or something :-)

On a serious note, it is rather odd. Running the company means I can, with ease, take a day or a week off and do whatever I like. I have a lot of very competent staff that run things for me. If they really are stuck for a decision they can always call me. I don't take much time off. Some times I'll go home early and watch TV or more likely work from home. But in general it is my choice, and I choose to work. So the idea of a sort of enforced holiday is strange. I feel like Sheldon forced to take his annual leave.

2014-05-03

Virgin: The plot thickens

I do mean the plot don't I? Not sure.

After they wrote to me and stated:-

Just to inform you we provide copper coax cable till the main box and from there we use the fibre optic cable to install the services at the customers property.

I wrote back:-

I am still slightly puzzled. You are stating, quite clearly, that you "use the fibre optic cable to install the services at the customers property". This is what I expected for "fibre optic broadband" ordered at my home. However, I watched the installation and have looked at the cable used, and I can be 100% sure that you did not "use fibre optic cable to install the services at the customers property". Please let me know when you will be back to "use fibre optic cable to install the services at the customers property" as it is clear from your letter that this is what should have happened. Let me know if I need to take more time off work.

I also asked about the install being delayed and who they use for ADR. To my surprise their reply today, from someone else in Virgin, is :-

Just to inform you we provide copper coax cable till the main box and from there we use the fibre optic cable to install the services at the customer’s property.

They then invite me to call them if I want to discuss further. They did not say who they use for ADR.

This seems to suggest that rather badly worded phrase is in fact a stock answer.

This could get fun!

2014-05-01

BT vogons

I would not normally go in to figures as they are commercially sensitive, so sorry if I am a tad vague in this post, but the latest stupidity from our favourite telco just left me feeling I wanted to blog it.

As a company we file annual accounts, but as a small company these are abbreviated even though we turn over many millions a year. This means that credit reference companies have very little to go on in terms of recommending a credit limit for us. This is no surprise.

It is just as well as we rarely want credit from our suppliers. In the case of BT, our biggest supplier, we have 28 days credit to pay the bill, most of which is for services in advance so we are paying at the end of the month that the service for which we are paying is actually provided (hardly "credit" in any normal sense). Given the significant time it takes to check and dispute the BT bills every month, 28 days is more than reasonable. We always pay BT on time, every month, like any supplier.

We order stuff from BT all the time - we order many new DSL and FTTC lines every day, and we order fibre circuits for customers (Ethernet) regularly. We order more bandwidth to BT every few weeks. We pay BT hundreds of thousands of pounds every single month, on time.

But we recently ordered a new host link from BT. This is increasing the capacity we have from two pairs of gig links to three. It costs around £10k to install and £2k a month extra with a one year commit.

Given the amount we spend with BT and the fact that we regularly order an extra 50Mb/s on the existing host links (which costs a lot more than £2k/month extra), with no problem, and given that we have paid BT of the order of £10 million over recent years, on time, we did not expect a problem with such an order - neither did our BT account manager. Why would we?

Well, for some reason, some pen pusher in BT decides that the policy of credit vetting a new host link order applies and he checks our credit rating. Unsurprisingly D&B think we have a low recommended monthly credit limit (I think they said £7k), so we fail the credit check for a new host link! This is somewhat crazy, because, BT plc are the one company that has a really good idea of how much we do - they know how much we pay them, on time, every month. They can see how many decades we have been doing that and can see how we are growing steadily. If anyone was to provide a credit reference for us, it would be BT plc! But they ignore all of that and decide that we are not a good credit risk!

To add insult to injury they decide that this risk only applies to the new host link order. That we can carry on with all of the ongoing business we have, hundreds of thousands of pounds a month, no problem. Even the suggestion that we start paying by Direct Debit for most of it was happily withdrawn when we pointed out that not one of BT's bills has ever been correct so DD would be bad for BT.

They decide that the host link order requires not only paying the £10k install up front (which, I don't have a real problem with), but that we are such a bad risk and might not pay the £2k/month, so we have to lend BT plc £6k (three months worth) for 12 months (and then they repay it with interest). I.e. they are worried we won't pay the £2k for the third host link but the remaining several hundred thousand pounds each month is not an issue.

[I did threaten to take them at their word, which is that the deposit is "for 12 consecutive payments", so I was going to pay £2k each day over 12 days and then demand the deposit back - that would have been a fun argument, but we pay them so much a month we can simple withhold the amount we are arguing about!]

It worries me that there are people in BT plc that are so blinkered that they think like this, that they think we are a bad credit risk at all, and then only for this one specific order and nothing else.

It smacks the poor account manager in the face as he has been pushing for us to buy more from BT, expand our business, and so on. He has bent over backwards to sort the order for the new host link, and tried to point out how stupid this nonsense is, but failed.

I was sourly tempted to video taking £16k in cash to 81 Newgate Street and dump on the desk, but having had a few days to think, we are actually putting off the new host link order for now. We were being very cautious anyway and considering lead times, but in practice we have a simple alternative, which is to put more new lines with Talk Talk rather than BT. That solves the host link issue for a long time, and gets us better deals with Talk Talk. It is a blow for our BT account manager, and I really hope he kicks up a hell of a stink in BT over this.

The plan now is that we look at a new host link a bit later in the year. If ever we find usage and ordering changes to need one sooner, then we can jump through BT's hoops if they really want (and maybe will video handing over cash), but for now we do have plenty of head room. We'll certainly not compromise quality of service to our customers over something as silly as this. The money is not the issue, it is the principle.

[mind you, we did wonder what rate of interest they would pay - maybe paying BT some up front as a deposit is a good thing if the interest they pay is better than the bank!]

As a BT shareholder it worries me the way they do things. If they really thought we are a credit risk (which is crazy, and they know it) they should not be just making the host link order difficult but reviewing all that we do with BT, total credit risk, and so on. They don't do that. Now, if some other BT plc customer is a real credit risk we know that BT will not check anything until they order a new host link, and then only focus on that, which is easy to accommodate. That is a credit control department not doing its job either way. That is worrying. I may even raise a complaint as a BT Group plc shareholder.

Thankfully Talk Talk seem much more sensible. I can't recall, but we may have had to show them full accounts when we started, but they are more than happy now, and when their account team try and sell us more, their credit control people don't screw that up by trying to upset us!

Hard to come to a clean conclusion on a post like this - BT credit control seem to me to be Vogons.

BT advert

So, my latest ASA complaint, which they will no doubt ignore:

Advert states verbally "*no-one* gives you a more reliable wireless signal than BT" and shows on screen "UNBEATABLE WIRELESS SIGNAL".

This is factually untrue as other ISPs such as my employer (Andrews & Arnold Ltd) offer broadband packages with a three pack of Unifi WiFi access points. By using three access points the signal is more reliable over the area of a home or office as devices can switch seamlessly between them to maintain the best signal. This means other providers are offering a more reliable wireless signal than BT.

My issue here really is that the advert makes no sense. They are selling a "broadband" service, but selling based solely on the specific type of device they happen to supply on the end of that service. There is nothing stopping anyone buying access points to ensure good wireless signal in their home/office completely independently of their broadband service. It is like BBC1 advertising the biggest TV sets or some such.

If you can't sell a service based on the merits of the actual service itself, then you have lost the plot I think.

Missing unix/linux/posix file open option

What I would like is a file open option for "create replacement file". The idea is that this makes a new inode in the same mount p...