Online and epos technology

[rovergull]

Hi Guys

I have been watching a few posts about problems with ordering from certain suppliers and noted comments with regards to what is wanted when ordering and what they think we should be doing when putting these systems in place.

Some of my experiences may well help give an idea into what goes into these systems and what it costs. I will also comment on watching one or two other web sites and what hey look like at first glance from my perspective.

As this is likely to be a fairly long topic I will probably split it into more than two posts.

 

I too got sucked into this "You have to have a web site" thing a few years ago and thought about it for a while. I investigated the cost of building one as I have no experience of these things and was instantly put off by the quoted costs, Nothing happened for quite a while until I was talking to a football supporter friend of mine who suggested couple of his work colleagues who had set up a company doing commercial web sites.

 

I asked them for a quote and they gave a fairly low one as they were new to the scene and I thought that if I could have a fairly simple and basic starter one, it would help me get off the ground.

I sent them some ideas and got a price back of about £1400 with me having to also pay to sort out fees for the domain of £250 it seemed all set.

Then in 2005 after agreeing what was needed of the site we were online. I hadn't gone more than a week before I realised the problems with the site. It just didn't work properly. Now there are probably hundreds of you out there champing at the bit who are thinking ah but you didn't ask me. Well yes we did! in the run up to appointing these guys to build the site we asked several who said they did this sort of thing for a living and got rough quotes of £2 to £3k so that's why we didn't go ahead.

The web site took some fixing and after several more weeks we finally got going. More bugs then appeared and having told them to fix it we were unpleasantly surprised to be given a quote of some £380 to fix a problem that they should have been aware of in the first place.

What and where do we go from here. well we mulled over the problems and realised that this bunch had gone as far as their limitations could take them and despite the problems we had to look elsewhere.

 

In discussing this with a friend up in Scotland he asked to take a look at the site. we duly gave him all the info we could and left him to it. After a while he came back to us and said it was fixable but some fairly basic mistakes had been made and it didn't help that for some reason the database was built with a protocol called ASP. Many of you experienced in these things will recognise this.

I asked for quotes and Paddy came back and said his guy didn't know anything about ASP but the good news was that he did. and so He fixed it. Oh boy did he fix it. The site worked brilliantly and he added loads of goodies and extras to make it look as professional as possible. I thought it a good idea just to outline what i wanted it to do but let him have a free reign. Now this guy is brilliant and he did it for the love of it. I did other services back as a thank you as a way of paying him back, I didn't have anywhere near what it would have cost if done by a professional company. Thank you Paddy and see you soon up in Edinbourgh.

 

The site carried on for a while until some things started to malfunction. and then it started to occur to me how much work was needed to just keep up with the site and keep it updated.

If a problem occurred it could have been the hosts who sometimes changed things on the server. More time spent trying to find the problem we didn't know about and they forgot to tell us. ( we have now changed hosts as a result)

We had some basic problems with the site not adding up properly. Not taking customers to the right page and all sorts of other annoying problems.

it all takes time to sort out. I also became aware exactly how much time it took to enter. Each item had to be entered singly. I now know it is possible to have items entered in batches but this wasn't available to us at the time. I wasn't fully aware what happens when you uploaded photographs. the file size soon filled up our database and I found I couldn't load any more products. This meant I had to bring on another program and resise loads of pictures. So to start from stage 1 meant either getting discs from manufacturers and then resise hundreds of pictures or if not available then take photographs and then resise them, once that is completed we can upload them. This process takes 5 mins per item if taken off a disc or up to 15 mins if I have to take the picture myself setting up the product getting light right and then taking 2 or 3 pictures to see which is the best one. (15 mins also included setting up the studio).

All the while we tried negotiating with several carriers to get much better prices for carriage if our goods. More time spent not selling and much wasted as most set up appointments but didn't keep them. I even went out to one or two and they couldn't even be bothered to stay in for us or tell us they were going out.

We encountered various other problems over the next couple of years but eventually got the site working o.k. It would be another year before we entered stage 2 and that the next story available later

Rovergull

[Kenton]

Sadly a tale of many small business and even mediums sized business and government departments. They do not set the project scope at the start, often not really their fault as they view it from a total bewilderment of what is required what can be done, the time it takes and options available. In fact there are many software development project managers out there with little more knowlegde or a blinkered approach.

 

Probably the first thing to be aware of is that most professional software developers have an hourly rate of £60 - £100 per hour. If you are running a from scratch project you will need one not a code monkey from the third world. So you can soon see the £££s vanish unless you are buying an off-the shelf product. As we have seen with IPB that gives you no guarantee it will work correctly and even less of a guarantee that it will do what you want in any way like the way you want it.

 

fees for the domain of £250

That is an astounding rip off. You can get a domain name for about £3/2years (.co.uk) £10/2years (.com) and most.

The domain should be separated from the hosting costs as the domain is a capital item - the hosting is a running expense and you will probably be billed monthly. However about £50/month brings you a server with high spec database and lots of other features including more memory than you are going to require unless you are Hattons.

 

More bugs then appeared and having told them to fix it we were unpleasantly surprised to be given a quote of some £380 to fix a problem that they should have been aware of in the first place.

If you have signed off the project as "fully functioning" you can expect to have to pay for changing features. At the end of the day all this comes down to the contract. On going support, help desk, remedial work. What the customer often sees as a bug many times is just that they are not doing things the way the system was designed, and that comes back to the original specification and contract. No professional software developer is going to sign up to that sort of open-ended support.

 

After a while he came back to us and said it was fixable but some fairly basic mistakes had been made and it didn't help that for some reason the database was built with a protocol called ASP.

Oh dear, more lack of knowledge. I hope he really meant that the site was written in ASP (becasue if he does mean the database is ASP built then you have more problems)

Your database should be one of the more common and well supported DBMS (typically MySQL, MSSQL, Oracle .. there are a few others)

The language that drives the web pages should be either ASP or PHP (simplest options) - there is nothing wrong with ASP other than it commits you to the Microsoft family (PHP commits you to the Linux family) All this and its important implications should have been communicated to you. It should have been your choice, it is your web site. The selection should not have been a problem for the developer.

 

I also became aware exactly how much time it took to enter.

Something every user of any system fails to really comprehend.

.... even making posts to RMWeb blogs

 

getting discs from manufacturers and then resise hundreds of pictures or if not available then take photographs and then resise them, once that is completed we can upload them. This process takes 5 mins per item if taken off a disc or up to 15 mins if I have to take the picture myself setting up the product getting light right and then taking 2 or 3 pictures to see which is the best one. (15 mins also included setting up the studio).

optimising and resizing should be done by the webserver, batch upload by ftp, If you are taking the photo then also you should be doing that in batches but there is no avoiding that one off hit - and a reason why so many small sites have few product photos.

[beast66606]

Only a couple of things I would disagree with ..

 

It should have been your choice, it is your web site. The selection should not have been a problem for the developer.

 

The developer should not be expected to understand every technology, but if they did they should recommend one with reasons, as a non technical person the OP would not be in a position to judge which technology is best and if they don't they should say "We don't understand ASP (Active Server Pages)" or whatever "so we can't offer that as a solution"

 

optimising and resizing should be done by the webserver,

 

Mmm - maybe, but it might be better to resize and optimise on a local PC and then upload finished images, saves the load on the server. 5 mins per image off a disk is about 4mins 59s too much !

[Kenton]

The developer should not be expected to understand every technology, but if they did they should recommend one with reasons, as a non technical person the OP would not be in a position to judge which technology is best and if they don't they should say "We don't understand ASP (Active Server Pages)" or whatever "so we can't offer that as a solution"

If any software house I came across was unable to explain to me VERY good reasons why they could not use one technology over another they would not be contracted for the job. I can understand the developer having a preference but not forcing it on the customer. However, in this case it does not seem to have had a developer or any project management but has been partially produced by novice code monkeys. The overall cost explains a lot. In software development you get what you pay for as in most fields of expertise.

 

Image optimisation and resizing presents very little load - it should after all be a one off hit with the actual final images stored static as files. Loading the images into a database and converting and resizing on demand by the client makes little or no sense. All that extra processing simply to get them out of the db so the webserver can send them back to the client as files.

 

This site also will have a relatively small number of visitors and activity.

 

The problem with having the user do all the file manipulation is as we have seen takes time they don't have, and leaves it open to that most common bug - user error - Computers are at their best automating processes according to a predefined set of rules (as in "if xyz do something the user might otherwise forget/change their mind on/think about/...)

[beast66606]

If any software house I came across was unable to explain to me VERY good reasons why they could not use one technology over another they would not be contracted for the job.

 

I assume you know every single language, technology, database, etc. etc. then ? - or whatever is the equivalent for your job ? - strangely most of us are human and don't know everything about everything otherwise it would be a boring world.

 

The garage I take my car to fixes it quite happily but I know they don't fix all other makes of cars, however that doesn't stop them doing an excellent job on mine - it would be stupid of me to take it elsewhere just because they don't fix other makes. If someone brings one of the makes they don't work on in they simply advise the person to go elsewhere (well to a specific place as they have a reciprocal arrangement) - simples.

 

"who had set up a company doing commercial web sites." - doesnt mean software house, I know someone who sets up websites but I couldn't discuss operator overloading in Pascal vs C++ with them, or why lots of C programmers are lazy and don't do garbage collection which results in poor coding and memory leaks, we've also never discussed managed code vs unmanaged code but I'm looking forward to that discussion ...

 

However, in this case it does not seem to have had a developer or any project management but has been partially produced by novice code monkeys. The overall cost explains a lot. In software development you get what you pay for as in most fields of expertise.

 

Bit harsh and if you want my time you would pay £750* per day (assuming you get the discount), a project manager would be a similar rate so if £1500 per day is ok then he got - 2 days work (with discount), so maybe thats why he had so many problems ?

 

Calling anyone a code monkey is just an insult.

 

Image optimisation and resizing presents very little load - it should after all be a one off hit with the actual final images stored static as files. Loading the images into a database and converting and resizing on demand by the client makes little or no sense. All that extra processing simply to get them out of the db so the webserver can send them back to the client as files.

 

 

Actually image optimisation and resizing IS a big hit, it uses a lot of CPU - but then I guess with only 350,000 photos (digital) I will be just guessing on that "fact"

No one suggested client side processing, if you read carefully I suggested resizing before uploading to the database rather than after downloading as you suggest I suggested.

 

This site also will have a relatively small number of visitors and activity.

 

Based on what information ? - he could be selling something that lots of people want, or he may have an unfortunate reference which means a lot of hits, from disappointed people - are Hattons (other small sites are available) a small site with few visitors and activity ?

 

The problem with having the user do all the file manipulation is as we have seen takes time they don't have, and leaves it open to that most common bug - user error - Computers are at their best automating processes according to a predefined set of rules (as in "if xyz do something the user might otherwise forget/change their mind on/think about/..

 

Quite right but again you have assumed that I meant everything had to be done manually, I never mentioned anything about how I would do it a lot quicker as the OP didn't ask.

 

 

 

* but I can program in lots of languages, technologies, databases, integration tools etc. etc. so I'm quite cheap for my skill set but unless you work for the NHS you won't get me.

[The Stationmaster]

Speaking from the viewpoint of a (past) buyer of both systems and software I think one thing which must be got absolutely right from the very beginning is 'what I want it to do'.

 

Provided the semi computer literate buyer (i.e he knows how to get into WORD or send an email but that's about it) and the technical expert seller can communicate the problem is part solved.

 

Back in the late '90s I was involved with some previously ordered software 'in development' (overdue and costs escalating faster every day of course) so just started asking idiot questions - the real basics. Net result - we stopped the latter part of the development because the software house said they couldn't achieve it. Reason - initial spec not properly explained or understood and programmers promising the moon.

 

I then traipsed round Europes (mainly electronically alas) looking for a replacement, my successor took up the hunt after me. But every time the panacea was presented to us (usually already in use in the airline industry, occasionally in railway use but 'still under development') our simple set of questions never got the right answer - it could not do what we wanted in the way we needed the output and we couldn't change all of our spec to suit what the system could do, or afford to pay for it to be rewritten to meet our spec. So I didn't buy and my successor still hasn't.

 

The buyer must get the spec right and get it complete (possibly with some IT professional help) because otherwise he will never get what he wants. And the developer must deliver what he says he can deliver. Simples

[beast66606]

In my 20 years+ experience the problems are most frequently caused by

 

1. - The buyer not specifying exactly (or even correctly !) what is required

2. - The buyer wanting more than was specified - for free, "creeping elegance" we call it

3. - The buyer wanting it quicker than was agreed and often with creep.

4. - The buyer not testing as things progress and then commenting too late in the development cycle that it doesn't do what was required (see 1)

5. - No communication between the parties.

6. - Bugs.

7. - Developers not understanding the buyer requirements (not the same as the buyer not specifying correctly)

8. - Technology

 

I guess your experience (from the other side) is very similar SM !

[Kenton]

I assume you know every single language, technology, database, etc. etc. then ? - or whatever is the equivalent for your job ? - strangely most of us are human and don't know everything about everything otherwise it would be a boring world.

Having run my own software business I can almost answer that in the affirmative - I could not claim to be a specialist in all languages or all technology but any customer coming to us would have been given a pretty wide choice of the then current offering. Conversant in just about everything from C,BASIC,PASCAL,FORTRAN/(X)HTML(CSS),ASP,PHP/Java,Javascript/MSSQL,MySQL,Oracle and even some machine level work we could offer a pretty wide range. But when it comes to coding there was nothing quite like a cheap graduate of the Indian sub-continent.

 

But sadly, or perhaps happily, as you know technology both hardware and software changes rapidly that it is impossible to keep up. But collaboration or buying in expertise in other areas is not unknown. Gladly I have moved on to other things.

 

Dealing with the techie-user interface is always the most challenging. Especially on large projects where even the user doesn't understand the processes involved accross multiple functions.

 

Your analogy with the garage doesn't quite work - as the customer was looking for a website design. Not specifically for a particular model of website as in specifically a ASP/MSSQL coded running on a MS Platform.

 

"who had set up a company doing commercial web sites." - doesnt mean software house

Sad, but true. But then they should be using off-the shelf package rather than specific bespoke work.

Most such "Web site builders" I consider should be handling the GUI rather than the behind the scenes coding or operational design. The look and feel (nice and appealing appearance) is a different world.

 

Bit harsh and if you want my time you would pay £750* per day (assuming you get the discount), a project manager would be a similar rate so if £1500 per day is ok then he got - 2 days work (with discount), so maybe thats why he had so many problems ?

We agree there, and my costs were based on what I was paying developers in 1999. No reflection on the OP at all because every one would approach this from the point of not understanding the time and expertise (therefore cost) actually involved.

 

Actually image optimisation and resizing IS a big hit, it uses a lot of CPU - but then I guess with only 350,000 photos (digital) I will be just guessing on that "fact"

No one suggested client side processing, if you read carefully I suggested resizing before uploading to the database rather than after downloading as you suggest I suggested.

Firstly, the number of product images used by the OP's web site will I suspect fall a bit short of 35000 but again the storage requirement for those images should have been part of the developer's brief along with expected/predicted number of those images displayed per page/visit. Certainly not all of them are going to be displayed by every page or even every visitor.

The point about resizing is that the owner/user should not be doing it. The user should just upload the image file. The webserver then takes a one off hit to resize and optimise the file. There was never an intended suggestion that this should be done visitor/client side. Of course you are correct that there is no reason why the optimisation program could not be run standalone on the user's PC prior to/even commensurate with uploading to the webserver. It would just not be my preferred approach as it potentially has support issues and puts the onus back on the user.

 

Fair to say my assumption was that the OP's site was "small" and that Hattons (et al) were bigger. But that is only based on the information that the whole project was undertaken on the "small" side. I suspect a "let's test this web thing out" rather than based on a clearly identified revenue stream.

[Kenton]

In my 20 years+ experience the problems are most frequently caused by

I'll add one more that I have found more common yet is even more dificult to solve.

 

The buyer/client not even understanding their own business. No process, no documentation.

 

But getting them to acknowledge this before 1. is sometimes difficult.

 

Though from the other side I acknowledge that the programming team do have a habit of introducing "that would be nice if" refinements that are well beyond the original spec all of which complicates the code (frequently also poorly documented) takes time and costs.

 

That is why good project managers earn what they do.

[The Stationmaster]

In my 20 years+ experience the problems are most frequently caused by

 

1. - The buyer not specifying exactly (or even correctly !) what is required

2. - The buyer wanting more than was specified - for free, "creeping elegance" we call it

3. - The buyer wanting it quicker than was agreed and often with creep.

4. - The buyer not testing as things progress and then commenting too late in the development cycle that it doesn't do what was required (see 1)

5. - No communication between the parties.

6. - Bugs.

7. - Developers not understanding the buyer requirements (not the same as the buyer not specifying correctly)

8. - Technology

 

I guess your experience (from the other side) is very similar SM !

 

The one thing we didn't have on the biggy was number 4 on your list, I was able to trace elements of all the others over the previous 3 years of development prior to recommending it was time to pull out.

 

I've also come across Kenton's point about the client not understanding their own business or requirement far too often. I first witnessed it when attending promotional demos for rostering systems where it was obvious that some other attendees were so blinded by the 'bells & whistles' they overlooked the need to question some very basic parts of the functionality, and didn't even understand quite why I should be questioning them (one who bought the software soon found out - and dumped it).

 

Far worse (and in consultancy mode, post full-time railway days) I found when Independent Safety Assessing some customer specs for (subsequently aborted) auto-route setting software that part of the client brief for safety validation was nonsense and had been written by someone who didn't understand what was involved and how it had to be managed, scary (and that's why there are ISAs nowadays).