Can Website Code Suddenly Malfunction After More Than 1 Year

Is it operating with the same version of the browser which was originally used?

Yes, same browser, Safari, and same computer too.

The computer and browser are both fully updated.

"The computer and browser are both fully updated" - so has Safari been updated since the original installation?

I apologize, I should have stated before, but the malfunction is not on the viewers side, it is on the back end.

The site functions perfectly well to the viewer, ie, to my or any other browser.

The malfunction arises when I try to upload stuff in the back end, I now get an error message that says "Cannot access the database"

I've emailed the developer about this and about another lesser irritant and he doesn't respond.

His not responding has made me concerned as to why he would not respond and that has made me ask if code can be corrupted after more than one year, without an intervention of some sort.

Anything and everything can malfunction over a period of time. Have you tried removing the updates from both your computer and browser, back to the last time you noticed it was operating properly? And has any updates been done to the web site that you know of?

I am still on the exact same updates as I was at the time before the problem arose and I haven't done any updates to either my browser or to OSX since the time when the uploading was working perfectly.

It just came out of the blue one day when no other thing had changed, ie, no updates to my end at all.

This sounds to me like some array size of the code is now being exceeded. As you've indicated your uploading additional data.

You mean I could have run up against a limit on my allowable space allotment or something? That does sound like a possible cause.

Wouldn't the developer be able to fix that and charge me for it?

This is what's so confusing about the whole thing, I can't know what he is thinking or doing without a response.

Can you delete (make a copy first) some of your back side input? Then add new content to see if it's allowed.

If this works, then a simple memory declaration change should be doable in the code.

But without seeing the code, I'm just guessing. If this is the problem, it should be a very cheap fix to up the memory allocation size for the variables that are affected used to reserve space for your input.

Thanks for the suggestion.

I will try it and see what happens.

It's not all that unusual for a software writer to include a time bomb in the code. They often do this to extort more cash from you in the future. The fact that the developer is not responding could simply be that he is no longer around.

These time bombs are generally triggered by the computer clock reaching a certain predetermined date, and often the easiest method of ascertaining whether or not it is a time bomb is to alter the computer clock to a date sometime before the problems began to see if that overcomes them.

Sometimes a bomb is a one shot code whereupon once the date has been triggered a piece of operating code is deleted and cannot be restored by a backdate, other times it does work. If it does, it may just be a matter of keeping the clock backdated, but if that is not an option then you may be able to have someone check the code and remove the bomb, that is if they can crack the security.

It could be something as simple as a database that is now full. There is free databases around that are able to compete with the big commercial versions but not to the full extend. There might just be a limit of allowable space in the database or on the server or a bit of both.

You could try to talk to your web page host to find out. You could also try to delete old data and see if feeding new data works.

What actually is the application meant to do? Are you even collecting data?

The other option as already stated is that an update:

1. On your side (Browser, system, safety settings)

2. Web page host (availability of ports, Server update)

3. HTML rendering tools, scripts, dynamic content software update

4. or a mixture of the above

prevents the code to run properly.

Can you confirm it is unavailable for all users?

Pretty sure the developer could fix it. A good programmer also leaves documentation behind which could mean somebody else can fix it. Never seen this happening, but one never knows. If there is scripts involved they would be "plain" language and can be reviewed by another programmer.

Want to post a link so we can try?

Try the telephone, then.

'You or somebody else has done this operation too many times' type thang

You now have four OTs. You and that other guy who's not 'jdg.'

USB is actually on-topic. It's one of the most fashionable ways to transfer malware.

Unapproved USB devices, even power chargers, have been banned from US Dept of Defense computers since 2009. 2009!

And rightly so, as it was just about when* they** installed that 'stuxnet' in Iran, right ?

* Some say it was back in 2007

**Allegedly NSA or Mossad or one of those.

Besides, most PC motherboards will allow USB to serve as a boot-source, able to load a "fenced" TSR (in a protected area on RAM) prior to loading the operating-system, leave a segment-anchor on the hard-drive, and once the system loading is complete, thread or weave itself into a hard-drive, for permanent-loading with each boot in the future, thus marking itself as a part of the operating system.

Do you have the original code backed up somewhere? If you do, compare them. I would not mention your suspicion that the site developer changed the code surreptitiously until I had concrete evidence in-hand. They could sue for libel if it were made public but untrue.

I don't know where or how I could obtain the code except from the developer who is not responding, so I don't have any backups and consequently I can't compare them.

The reason I'm asking here is because I don't want to accuse the developer of anything unless and until I know that code can't be corrupted by itself without some sort of intervention.

Good call.

Web is not what it used to be. Javascript, CHS, PHP, Pearl and other web based languages are enough sofisticated to do just about anything including time related behavior change, exploit backdoors and other nasty staff if set to do that. But I think this is NOT your case. Something broke in YOUR system while "updating" it, it's not the code itseld, since you haven't touched it. Hypertext is mostly "calls" to those intepreter functions inside the "updated" libraries. Now even reverting back to old configuration may not fix things, since also settings may not be 100% compatible. Another possibility is something is messed on your ISP side. Not that uncommon, they "update", too. S.M.

The was an incident here in the UK many many years ago in the days of the first digital burglar alarms where the unit failed a month or so outside the warranty period.

The (simple) LCD screen displayed an error code that pointed to a message in the handbook that gave a phone number for help.

The help was free in the sense there was no call-out charge to diagnose faults, but if outside the warranty period, a free estimate would be given for repairs that would have to be paid for.

As you might guess, the unit could not be made to work on site because the part for the suspected (very rare) fault was not carried as a routine spare. The unit would have to be taken away (for their experts to look at) leaving you without burglar protection for a fortnight or more (they were very busy fitting new alarms due to the increase in burlaries in the area) but you could pay extra to jump the queue for an express repair.

Alternatively you could have a reconditioned unit at cost, or for a small extra fee, a brand new unit (that the engineer happened to have with him) fitted without delay and as a special offer (for that day only) they would give a big allowance for the faulty unit, and for an extra fee, an extended warranty, that would be included free if you paid extra for the engineer to fit the new unit at that visit.

A brand new alarm, an extended warranty, a discount, a part exchange allowance - and fitted without delay - what a bargain!!!

However, word got round the trade that there were a high number of failures for this particular make of alarm panel, which oddly only seemed to apply to those installed by one particular contractor, where over time, the failure seemed to happen at exactly the same period of time after the expiry of the warranty.

To cut a long story short, it turned out that the installer had found a way to make the internal clock, after a predetermined time, to flag up an error message.

It actually became a criminal issue and the installer was eventually caught, not so much by customers complaining as such, but more by the trade crying foul giving the genuine product and manufacturer a bad name. The installer was charged but I do not recall what the penalty was.

Part of the defence was that customer had been given good value for money. They had not been defrauded because the old unit (albeit with an error display) still worked perfectly well, and brand new replacement units, if purchased were supplied and installed at bargain rates - which were far less than the cost of their competitors, or other service contracts.

In fact everything said above seems to be fairly standard practice nowadays - with any machine - except you pay through the nose for repairs - the only difference then was the alarm unit being deliberately programmed to 'fail' just outside the warranty period.

Have you defragmented the hard drive yet? It may be that the code has become fragmented.

Updates can break things. A few weeks ago I read that the database commands that I use over and over are "depreciated" and will go away in the next version of the data base. I need to go through my code and rewrite every one of these calls. I have no way of knowing if this is what happened to you but it is possible for "good updates" to do bad things to code.

"...there's no such thing as perfect software, but that's not to say you're seeing symptoms of bugs..."

More likely, there's no such thing as "complete debugging".

The number of possible machine-states created by an application is astounding, compared to to the few points-to-be-achieved designed by the programmer, who can hardly ever foresee all the logical consequences possible, multiplying with each new command, function and delimiter, as they fill the lines of a program - Any program.

Any rare machine-state, out of the reach of the app user, is a potential bug.

Statistically, it's a miracle that software somehow manages to work, sometimes flawlessly..

I've listened to everything here and I'm taking your advice by hiring a Systems Administrator to look at it and hopefully keep it trouble free.

I'm negotiating with a prospective hire right now and he has suggested a good idea is to create a mirror site where all the problem solving and development can be done before introducing modifications to the real site which sounds like a reasonable idea to me.

One thing he has asked me for is permission to put his company name in page templates and code comments.

I presume these are adverts for his company and he has told me that these will not be visible to the public, only to those who look at the source code but as I do not know anything about code I am wondering whether or not it is a good idea for me to allow this as I would be relying on his word that it can not affect the site's performance in any way.

Can you offer me any insight or advice on this request of his, more exactly, as to whether or not it is safe to allow this?

It's always a good idea to have a contract stipulating rights to the code. If s/he puts copyrights or proprietary statements into the comments, that could create legal hassles later. Claims of authorship/ownership in comments are customary and technologically harmless unless they are done wrong (quite unlikely and easily identified: the code doesn't run or a page looks obviously malformed). Your web site code probably already contains such from your previous developer.

Plan to require the services of a sysad periodically. You should be getting some of this value from your hosting service, unless you run a self-hosted site. (In today's world, self-hosting is risky due to the threats, required maintenance and potential liability. Look into a cloud subscription, whose fees should provide basic protections and coverage.)

Thank you for your advice, much appreciated.

The arrangement I am negotiating is for administration as and when problems occur, or did you mean ask for say a monthly check-up or some other appropriate time interval once the problems are rectified?

By self-hosted, do you mean hosted by the developer?

My site is hosted by the original developer so is that a cause for concern? The current contract ends December 31st.

It's my understanding from my prospective Sysad that any host will have access to the code so nowhere is guaranteed safe and he suggested staying with the existing host on the grounds that the developer / host is probably not interfering and the cause of the problem is most likely elsewhere.

One of my basic requirements from the outset was very fast loading of my front page and I thought that a cloud service would hinder that so I've not considered it until you mentioned it.

Am I wrong to assume that a cloud service will slow the front page loading time?

Remember the caution about "free advice." :) The last time I tried to run a web site was in 2004, but it was one I wrote from scratch. I'm sure there is more up-to-date experience on this forum, if not then "out there."

Self-hosting to me means someone running the server out of their bedroom or back office, which is dangerous and time consuming.

Someone needs to patch the system and respond to immediate security threats that are identified. But, someone patching the system could easily check your log files periodically and respond to error messages immediately (by calling you). Perhaps the current host has a next level of service you can sign up for.

Thank you very much, much appreciated.

"Or is it actually feasible that code can suddenly malfunction after more than one year of operating successfully as it was originally specified?"

Yes. Very. The conditions under which the programme is operating are constantly changing. Its environment is changing through software and hardware updates. The data running through the code is changing.The total amount of data in the database is changing. The date and even time is changing. Access to the system is changing through browser and OS updates. Bugs found months or years after the system was first launched are common.

Try every way to contact the developer. E-mail, phone, text, write.... drop by. Remember you signed this code off and unless you can prove it was dodgy the developer can quite reasonably ask for a fee for writing updates. If you can't find the developer then you'll need someone else to fix it. First step: getting a copy of the code. If you can't find the developer find the company who is hosting your code and ask them for a copy. Expect bureaucracy and lots of layers of proof. HOPE the code has been written in a comprehensible form with lots of notes and explanations.

The big lessons here are: always keep a copy of the original; and, every programme has a bug. It's just a matter of time.

"...about 15 months after the delivery date the site is starting to show errors and the developer is not responding to my questions as to why it is happening..."

Many freelance or hired site builders leave several "Back-Doors" in their source code, in case the site will generate wealth for it's owners, while the hard working programmer was left out in the cold.

Some, with hidden agenda or with failed expectation of payment, leave a "Time-Bomb" in their source code.