Forum "Who's Online" is broken

[#1]

---

@greygit1 wrote:

I never knew there was a time period that could beseen via a click-through. It wasn't immediately obvious. You've obviously been working with this longer than I have, and I was obviously falsely assuming it was a cumulative number.

Are there any descriptions/documentation to prevent future similar mistakes?

---

To get that screenshot -if it was yours - you would have to have clicked the 'top fix authors' tab, and then selected the 'last week (7 days) option.

No, no documentation, but you clearly either don't look at what you are doing,or are just being intentionally obtuse.

[#2]

I thank you for your tolerance/patience which I fear I may have exhausted.

[#3]

I note with interest that @NickBS was on here yesterday (or at least logged in at some point), but failed to respond to any of the recent relevant posts.

[#4]

I'm sure I've mentioned this before, but for some reason - either intentional or accidental, exactly an hour after a  user has logged in, they disappear from the 'users online' notification at the bottom of the screen.

My suspicious mind says this is intentional, to give the impression that the the forum has few visitors and allow BT Consumer a reason to close the it. I doubt there will be either a confirmation of this, as it would expose the plan.

[#5]

@jab1     

OK... so,

WHERE AM I?   Seems like I am "ghosting" somewhere ..... 

***************************************************************************************************************************

********************************************************************************************

Been on line since about lunchtime... NOTE THE TIME at bottom right corner.

***************************************************************************************************************************************

Also this is WRONG too...

***************************************************************************************************************

As can be seen by this post in General Chat 

**************************************************************************************************************************************

Definitely something weird going on with all those contradictions .....    

[#6]

Lunchtime to 21:45 >* 1 hour which fits with @jab1s theory.

Posts made in General Chat haven't shown in 'Latest Posts by Username' for what must be a couple of years since the forum was tinkered with during the time General Chat was closed.

* is greater than

[#7]

@shutter If you've read my recent posts on here, you'll know my thoughts.

@RobPN although I did spot the 'nonsense' post in the 'unanswered' list a couple of times just after shutter posted it, before it got any replies.

As I've said, someone is messing with the coding on this site, for reasons so far not admitted.

[#8]

I suggest that there is no conspiracy here. Just very poor quality coding or inconsistency across the platform relating to a user’s status.

We have seen evidence of connectivity changing target servers. We have empirical evidence of the online status being lost after an hour of being idle. I suspect that the management of “stay connected” is broken such that one does stay logged in but some parts of the service thinks you’ve timed out after the default 1 hour. I have never seen an idle duration on the detailed who is on line list greater than 60.


Such will clearly lead to defective usage stats however I similarly think that those in the pickle being discussed here, WILL be included in the GUESTS count - that is people accessing the service without being logged on.

[#9]

---

@jab1 wrote:

@RobPN although I did spot the 'nonsense' post in the 'unanswered' list a couple of times just after shutter posted it, before it got any replies.

---

Yes @jab1, but that's not the section of the forum being referred to, and that's been the case since the time I mentioned above.

IIRC the tinkering was done to prevent 'unauthorised' users (i.e. non-logged-in persons or those who hadn't qualified to post in GC, also search engines etc.) from even being aware of GC.  For those who fall into the categories mentioned inside the parentheses that is still the case, and they wouldn't have seen what you refer to.

Try it next time you spot an unanswered GC post  listed, i.e. log out and refresh the page! 

[#10]

@Townman , @RobPN  OK. In reply to the user display issue - why has it recently (like since JonoH left) suddenly been really ropey - that suggests certain code has been rewritten -badly.

As regards the GC situation, TBH I'm not that bothered, it was only another 'new feature' issue that pops up on the odd occasion when a new topic is introduced there, and again is not consistent.

To be truthful, I think I may as well give up with this question, because I don't believe what I am seeing.

I admit to not being a computer programmer, but I have been closely involved with IT departments in the implementation of four different intranet management information systems, and I do know that to change what appears on a users display requires human intervention on the actual code - with appropriate and effective change logs/procedures.

[#11]

OK - where am I, and it looks like the system can't count - again!

[#12]

Yes human intervention is required to make changes - both intended change and consequential unintended change to the user experience. I do not believe that the product owners have the skill or wit to deliver the shape of the brokenness we see here. This will be an unwitting consequence of poor testing of code changes.

Having moderate experience with managing software development unintentionally breaking complex code … can be all too easy. One of the biggest issues in testing change in complex systems is that test environments are rarely (in configuration) as complex as live systems. Typically a test environment will be a stripped down platform having only one instance of the various “bits”. For example it is unlikely that operation across two (supposedly synchronous) database instances will be fully configured in a test environment (it will be ASSUMED that the distributed database tools do their job) and a test system is unlikely to encounter volume loading which fully exercises load balancers across multiple web service instances.

I think there are bigger issues in respect of BT Consumer Division’s “interference” in Plusnet - it is more likely that the demise of Business and not providing residential users with VoIP are the consequences of their icy fingered grip around Plusnet’s throat.

As for @JonoH yes he did a super job around here before taking on a different role improving on line help. He still lurks around here. His replacement was more interested in EE and BT Retail … and their replacement has yet to fully show his colours. To date he has been more responsive that his predecessor (as shown by his presence here) … but has yet to achieve Jono’s effectiveness with the product supplier.

Frankly the performance of this platform continues to question the wisdom of the decision to choose it. Such is far from untypical of computer systems - the replacement looks better than what it replaced, but never functions as well. This arises from the decision makers (not being system engineers) using appearance to influence choice, not functionality. New systems might look “better” but are invariably less functional.

[#13]

Fair enough - I am now out of this topic - others may wish to pursue it, but I have now lost interest.

[#14]

Moderators Note

This topic has been locked, 

The issue may or may not be fixed at some point, however this thread is no longer surving a purpose.