The comment registration system here is still a PITA. I know; I get so many complaints from so many people, yet at the same time, I need the dog-damned thing in order to manage the horrendous pile of spam and troll-trash spilling over into the comments.
So I'm going to compromise a bit. I will occasionally switch off the comment registration requirement for random periods of time, just so people who are locked out by its clumsiness can get a word in; but I will also sporadically switch it back on whenever the noise gets to me. Which might be every day. Or every couple of days. Or every weekend. I don't know, it depends on how much annoyance I can handle.
There's a hierarchy of increasing pain at work here.
-
The best solution would be for SixApart to get in here and fix the registration system so it worked reliably for everyone, which I suspect would mean moving away from their beloved TypeKey/Moveable Type in-house schemes. Which suck.
-
Everyone who can use the registration system continues to do so, while I sometimes allow anonymous/unregistered comments for people who can't register. This sucks for the unregistered, because you never know when the registration lockout will come down, but at least they can comment sometimes.
-
I just switch on comment registration full time. People who can register feel little pain, people who can't will simply be silenced.
-
I shut off comment registration permanently. Everyone rejoices, except for me, who suffers horribly behind the scenes. In this non-egalitarian universe of Pharyngula, I'm sorry, but my pain counts far more than the pain of a million readers.
We've been operating under solution 3 this past week, and I'm switching now to solution 2 for the indefinite future. We will not be going back to solution 4 ever again, because it makes me cry like a little baby every morning when I wake up to the nightmare of cleaning up the comments here. Solution 1 is what we all dream of, but we are at the mercy of SixApart, and they are evil and capricious gods with their own agenda, which does not seem to involve enhancing the interactive part of blogging.
The bottom line is that if you can't get registered (it's not your fault, the blame lies on the weed-smoking brain-damaged monkeys they apparently hired to code that stuff), you still suffer, but I will intermittently switch off the hell-code so you can type in a few words. Then I'll slam it down again, and you won't know when.
- Log in to post comments
Hope this plan works for you. I have no intrinsig problem with registering, I just can't find a link that says "click here to register" or similar...
How *do* I register?
Come on, PZ, design and code your own squID-based commenting system - you know you want to!
If I had the power, yes, I would just bring in a simple captcha system which would clean up 99% of the problem. It would be easy, too.
Unfortunately, I do not have access to most of the software, and can't go in and tinker with the guts of the machine. Which is a good thing, since this platform is shared with about 80 other blogs, and if I could do it, so could they. It would be chaos!
Perhaps I will get to comment occasionally. (not that I've ever been missed)
I should also mention to all the people who have emailed complaints to me that, although I almost never reply (I can't, there are too many of you!), if your complaint contains any details at all about the problem, I hit the forward button and send it on to the poor suffering staff at Seed, who then feed it on to the deservedly suffering techies at SixApart.
Maybe someday they'll get the message that their fucking registration system is fucking broken.
help! never a big presence in comments, but pretty sure i've been unduly squashed.
I tried to register with MT twice, but never got the emails to confirm the reg. Luckily, I've found that signing in via livejournal works just fine - and it's pretty easy to set up a throwaway LJ account if you don't want to use a real one.
Octopus
I recent the suggestion that monkeys smoke weed (and I hate monkeys) - even to lessen the impact of Alzheimer and Parkinson.
No, Waldo, it's still swordfish. It's always swordfish.
Who do I need to e-mail to [HRRRNGRRR] politely ask for a Captcha?
Let's pharyngulate their inbox.
And the spamming starts up again.
This sounds like a reasonable compromise and I hope it works for you. This go-round the registration solutions that were successful before for me, failed miserably. I did not think that forwarding to you the gibberish error messages I received would be helpful, but now I know better.
Thanks!
Why this is so difficult for Seed/whomever is beyond me. Plenty of other blogs seem to moderate it automatically and reasonably fine, and yet this one inundates its prestigious authors with trying to weed through spam.
I can't BELIEVE, that after all this time SixApart (S I X A P A R T) STILL hasn't been able to fix something as trivial as a registration system.
ESPECIALLY when a famous blogger with a million (!!!) readers openly (and deservedly) calls them incompetent at every occasion he gets.
Not just only from a technical point of view, but from a business perspective: it's INCREDIBLE that SixApart lets this happen!
Can we get some SixApart exec explain to us why this is so damned hard to fix? It's not that they have nothing to lose.
What about anonymous comments with ratings? You have enough regulars and reasonable readers that trolls would get rated-away pretty quickly.
You don't have to shoulder all the work!
"The comment registration system here is still a PITA. I know; I get so many complaints from so many people"
Why not just get people to sign up at google?
I mean is it really that hard to type in www.google.com, sign up, then come back over here and login?
:/
Can you appoint moderators?
The registration was quick, simple and seems to be working for this newbie.
"I can't BELIEVE, that after all this time SixApart (S I X A P A R T) STILL hasn't been able to fix something as trivial as a registration system."
Because it works, and it works for most people.
Just so you know this is noname420, I just signed up then, took 5 seconds.
Well, maybe you don't have to shoulder the burden all by yourself. If you could find a handful of trusted individuals, perhaps they could spend some of their free time cleaning things up. I don't know if that's possible (permissions and all) but it might make things easier all around to share the work.
Consider it.
Maybe it is user error on the people trying to register? I had no problems registering. I just login with my Google Account (OpenId I suppose).
Another solution would be to find a student or two in the journalism school and press them into service cleaning up the comments. Online journalism is a study area by itself now, and what better experience than dealing with "the public"?
"Unfortunately, I do not have access to most of the software, and can't go in and tinker with the guts of the machine."
Actually you can:
http://www.movabletype.org/opensource/getting-the-source.html
It's open source.
Tinker all you want ;)
@ ikt #24:
It's the infernal depths of Seed's (ie ScienceBlogs' owners) servers and software with which PZ is not allowed to tamper. He's not on his own blog any more, ie the original pharyngula.
No I can't. This instantiation of the blogging software is a shared utility by a group of people. It has to be managed.
Does anyone have a link where I can attempt to sign up?
#17, I registered yesterday, hit the "re-send validate email" twice, and I still have not received anything.
These fuckers are like vultures, well, except that vulture do provide a useful, if rather distasteful, service.
And how does registration help? I mean help ma boab could come back under a new registration name, by using a different computer.
Christians are the worst, man. And WE are the ARROGANT ones. yeah, right. Most Atheists are very pleasant kind-hearted folk.
I'm not entirely sure, but I think that even if the sign-up borks, you might still be able to comment. It just never tells you that you can. That, or I've just been lucky enough to hit the site when the walls are down. I've *never* received the confirmation email, even after trying several times.
@sef
"The best solution would be for SixApart to get in here and fix the registration system so it worked reliably for everyone"
SixApart are the main coders for the software that is 'moveable type' however as the system is OPEN SOURCE, Mr. Myers can look at the code and contribute to it.
As per the above quote he is blaming SixApart for the problem, he is blaming their code, which as I have just stated, he is more than welcome to check out and 'fix' the issue himself.
If the registration system is broken because of something incorrectly configured by THIS server, then he should be blaming "infernal depths of Seed" administrators and talking to them, WHICH HE CAN.
And to those who have registration problems, SIGN UP AT GOOGLE, LOGIN HERE, AND THEN TELL US/SIXAPART/IDoS WHAT THE PROBLEM IS, BE SPECIFIC, USE SCREENSHOTS IF YOU CAN.
This problem is entirely solvable and I don't like hearing about how it doesn't work after it just did for me on multiple operating systems, on multiple browsers, on multiple machines, that little yapping dog in the back of every IT Admins ear is getting louder.
PEBCAK!
...well the ones who got to Atheism through Science-learnin', anyway. Not the ones who are "angry at God"
#20:
That should NOT be a reason to not fix it and leave those people for which it does NOT work (all the time) hanging!
(I'm one of the people who intermittently can't comment, no matter WHICH openId (I have various) I use! It's just flaky. That it works for you or even for MOST people, really doesn't help me a whole lot)
Why don't you just hire some undergrad to zap the spam every morning? They can do it right after they bring you coffee and massage your feet.
Signing in with my Google account (just Gmail, really), seemed to work fine. Had to allow Google cookies for the session before Google could reply to a request for account information from Scienceblogs, but otherwise, no problem.
FTR, running XP SP3, Firefox 3.5.5 with NoScript 1.9.9.22, Adblock Plus 1.2.2, Cookiesafe 3.0.5.
(And now I'll try previewing this...)
So far so good.
(Now Submit....)
#30: You can't stop the spam, but requiring registration makes it easier to slow down (registration takes time) and tidy up (when you find a spammer, you can find other spam from the same registered user).
@28,
then re-sign up on a different email address, clearly there is an issue between your email server and this one.
#39: I tried signing up for MT using two completely separate email addresses (home and work), neither has received confirmation emails despite two requests apiece.
I found that I could sign up using my Google account which works (as comment #37 indicated)
#33 (ikt)
Wow! Are you one of those support people who angrily respond to problem reports with:
"Works for me. Case closed (if you have any more problems: reformat your disk)".
MAN, I hate people like that!
You think all those people who complained about it are making things up? Or are too stupid to own a computer?
Which part of "it intermittently fails" didn't
you get?
Occasionally I have login difficulties with TypePad but really, it's not that bad. It seems to work just fine with Firefox even with NoScript fully functioning and ABP enabled. One has to give the required permissions to get everything functioning but once that is accomplished everything seems reasonably smooth.
If a developer can't reproduce a problem, he/she/it can't really fix it. You saying "it doesn't work at random" isn't really useful, as it could be broken for a wide number of reasons (you could for instance be using a shitty browser like IE6).
Anyhow screw MT, just sign-in using Google. Google always works. Sure you get an OpenID URL instead of a cool name, but you'll just have to choose between that, and not being able to post!
@40,
http://dl.dropbox.com/u/2161836/pictures/ok.png
I would recommend checking in with your system admin at work, just ask him to see if he's receiving the email, as my screenshot is showing, the system is clearly sending out emails, therefore the issue is with a system in between your email server and this one, hell it might just be your system admins spamlevel is 1 level to high.
@42
see #44
Tom @23 I don't know Tom. Getting students to volunteer as spam cleaners might be possible but journalism students? I think the task might be a little too demanding for someone whose best subject in high school was typing.
#44
That's the (bad) developer equivalent of the (bad) support guy's response "works for me, case closed". Equally closed-minded and equally unacceptable!
In my long career as a developer I have fixed COUNTLESS problems that I was not able to reproduce.
I didn't mean it to be useful! This is not SixApart's support blog! I was merely confirming that, yes, the system is STILL somewhat broken!
And while I personally have seen this with various browsers on various systems, an average reader SHOULD be able to register without a need for "multiple browsers, multiple operating systems, multiple machines" and a gazillion email addresses.
The whole point of PZ's post is: Something is Sometimes broken for Some people, and it causes a hell of a lot of trouble.
Is just needs to be fixed. Period.
Hey, just testing if this thing works...
#48
It did.
For you.
For now.
;-)
Uselessly_Verbose_URL (@44):
Google does not always work: I was unable to get in using my Google account. I eventually was able to register through MT, and now I'm able to sign in and comment with no trouble, but there was quite a bit of trouble along the road, and Google was by no means a foolproof out. In any case...
It's not a matter of being hung up on the coolness of one's display name: This is a community, in which people talk to each other. Commenter names that are literally incomprehensible utterly defeat the point of commenting in the first place.
I get that it's frustrating for developers to try to fix problems they can't replicate... but the fact that developers can't replicate the problem doesn't mean users aren't really having the problem. Your "aw, just suck it up, you lightweights" attitude is comprehensively worthless.
"I get that it's frustrating for developers to try to fix problems they can't replicate... but the fact that developers can't replicate the problem doesn't mean users aren't really having the problem."
What error were you getting when you were unable to login using your google account?
As a former programmer and current computer user, I can see the frustration on both sides. But ultimately, "right" comes down on the side of the user. If the software doesn't do what it's designed to do, or doesn't do it consistently, it's worthless.
I managed, yesterday, to get signed up through Typepad. But it was confusing, and the whole process (typepad and this blog) assumed a level of knowledge that many people don't have. Lack of nerdness doesn't make us unworthy of being commenters.
This is insane. We are in the 10th+ year of online communities. This problem should not exist.
@53,
no problem here :)
Signed up with 2 accounts with 2 different emails, signed in using google and openid, all systems are go here except for some reason the login page, not all the images are showing, which I'm just firing off an email about.
I'm really interested in the issues people have but I'm not getting any information other than "intermittently it doesn't work" and as it works for me, I'm not sure what I can do :s
"Artificial intelligence: how far are we from a system which can think and reason?"
This is at the right margin today. Don't I wish.
@52
"If the software doesn't do what it's designed to do, or doesn't do it consistently, it's worthless."
Oh yeah? How many people use windows?
Nuff said.
I am not signed in right now, but I do remember having issues with it in the past. There's a little checkbox that says something like "remember me" or "remember my password" ... and it never did. I just found it needlessly cumbersome.
The above comment was directed at (https://login.launchpad.net/+id/fGKTDPN) if that is your real name.
Anyway, another issue I have with this site is the ridiculous amount of time it takes to post a comment.
The Movable Type confirmation email will have the fields
if anyone wonders what to look for in their junk or spam folder (mine went right into the spam folder when I received it originally).
Heya this is "ikt", "test" and "https://login.launchpad.net/+id/fGKTDPN", I've changed my mind, the system admins at scienceblogs are fucking idiots.
Apologies for any blood levels raised :(
For what it's worth, here's what consistently happens to me:
I've successfully registered with MovableType. But when I try to sign in with that username and password, I get a message "invalid login". If I try to register again, it tells me I'm already registered. If I say I've forgotten my password, it promises to send me a reset email, which never arrives. Now you could say that my email server must have some kind of problem. But I get email from everyone else.
Another thing: there's no way during the login or registration process to inform whoever is running MovableType about any problems. I suppose it would be possible to search for those responsible. But why should you have to? A little user-friendliness here, please.
https://www.google.com/accounts/o8/id?id=AItOawnTAiIRbRIpbzIZTtwLDKEdcE…
Don't listen to them! I think that's a way cool handle.
Anyway, PZ's approach is right. Even if you can't reproduce the issue, you can still provide error messages and specific details about what you were doing. It's the next best thing to "come and get me the next time it happens."
Good morning all. Still lurking every day because registration hates me. Very happy that Cuttlefish has found the way to get through and congratulations to Lynna OM.
May I also wish everyone a Happy Holiday. Festivus is almost here.
I use the MovableType option, and it seems to be working NOW, but in the past when I did have problems it was always the same exact thing: A permissions error from the server. I thought it was telling me my password was wrong or my account wasn't privileged enough, but then I realized I'd seen that message before from webservers I've run myself. The "permission error" it's talking about is that the web server user (i.e. "apache") isn't allowed to get at the file it's trying to look at ON THE SERVER. Whether this is an excecute permission error (Trying to "run" a file that is flagged as being a passive data file rather than an executable program) or some other permission error, I don't know, but it's frustrating as all get-out.
One of the more common web sysadmin mistakes is testing changes to the system from a priviledged account rather than from an account that has the permissions of a normal user, and therefore never noticing the error when you set the permissions wrong (because it works from your own privileged account without complaint).
#60
You mean, as opposed to the SixApart folks?
Can you elaborate on this observation of yours?
Well, this was weeks (months?) ago, during the previous go-'round of Registration Hell®, so my memory isn't terribly precise on the point... but IIRC I didn't get any error message at all: I would attempt to create an OpenID using my existing Google account and nothing would happen. I was eventually able to create an OpenID through another route, but found the resulting display name intolerable, and I ended up going without commenting 'til the next time PZ turned registration off.
Then I tried to sign up through MT, and had a silly little dance with them because I had an old MT account whose password I couldn't remember: MT's promised password-reset e-mail never appeared — no, it's not in some SysAdmin's queue, because I used my private e-mail, and yes, I checked to see that it wasn't in my spam bucket, either on the e-mail server or in my local e-mail client — and MT wouldn't let me create a new account because there was already an account associated with the e-mail address that they couldn't successfully send e-mail to!
Finally I got MT to work, but I have no idea how: It's one of those cases where frustrated blundering and armwaving produces a completely unrepeatable solution to a completely inexplicable problem.
I know that's not very useful as Help Desk input, but guess what? This isn't the fucking Help Desk. People here are just commiserating with their fellow sufferers.
So if the problem is that a variety of users are having a variety of problems that they can't always describe and that the developers can't always replicate, but that don't go away... it seems to me that the obvious way to fix the system is to replace the unusable, unsupportable system with one that is usable and supportable. In this case, IMHO, chasing individual users' individual problems misses the point.
PS: MS WinXP Pro/IE7 at work; MacOS 10.4.latest/FireFox (continually updated) at home... FWIW.
Since it looks as though some actual coding/test bods may have shown up here, I'll refer them back to previous error reports on this recent thread (and there are remarkably similar reports on older threads from before the alleged fixing of the system).
You missed the other option from your list: Move Pharyngula elsewhere. Somewhere that can handle commenting. And captchas.
All true, but what a lot of users don't realize is that the programmer has to be able to reproduce the problem as the very first step to fixing it. If it never happens when the programmer is looking, there's absolutely nothing that can be done to repair it. It's not merely frustrating to try to fix a problem you can't replicate. It's literally impossible to, because here's the situation the developer is in:
"Did you see the problem before you implemented your change?" "No."
"Okay, did you see the problem after you implemented your change?" "No."
"So, did your change make a difference?" "How should I know?"
I suspect that the problem only occurs under larger loads with lots of users. So it ends up happening on the public site, but not on their playground testing area. To really reproduce conditions that appear on the live site, they need to spawn dummy users that are constantly hitting the login page from different test machines, all at once - like it would be in real usage.
I used to have a lot of trouble getting signed in. Then I noticed that you could use your Live Journal account to sign in and have had not a single problem since.
@65,
yeah, as on this page:
http://scienceblogs.com/channel/about.php#contact
Technical Inquirieswebmaster@scienceblogs.com
Sending an email to webmaster@scienceblogs.com results in:
Your message could not be delivered because of previous failures during delivery attempts to this mailing list. Please update the list with valid addresses.
Confirmed in the irc chan with other people just then.
If they can't even get a simple email system right.. FFS
what about installing TypePad AntiSpam ?
ikt@56 I'm not sure you actually proved your point about the worth of the Windows operating system. You have, however, made a reasonably accurate observation about the triumph of hope over experience.
I had problems getting in the at first. I had to create a whole new name with a new email address in order to post here. I do still have problems some times. Problems like this:
I enter my username and password. I submit these. When I submit them it behaves as if I have signed in. Under leave a comment it says "Thanks for signing in, Ol'Greg" as if everything is good. However when I post it tells me I need to sign in. We do this a few times. I try checking and unchecking the remember me box. Eventually I either get to stay signed in or I leave.
In my limited experience, the validation email is often sent to one's spam or junk folder. Check there.
As far as being able to sign in easily, yes that sometimes happens. In fact, I can sign in easily most of the time. However, even though I'm registered with typepad and have signed in easily hundreds of times, there's always going to be that one time when it just refuses to work. Case in point, December 12, past Saturday, when I could not sign in all morning and finally sent PZ an email with a copy of the coded gobbledygook I was receiving.
Steven (@69):
If "it" never happens when the programmer is looking, yet users keep reporting it, there are only two broad possibilities: [1] The users are either delusional or lying, or [2] the programmer isn't trying hard enough to duplicate the users' experience.
In an earlier comment the problem of doing development and/or troubleshooting from an account more privileged than the typical user's. It goes further than that: You have to duplicate the entire range of user experience, including the worst-case supported environment (just sneering at the user's "shitty browser" is unacceptable). They have to get outside what you refer to as their "playground area."
I face a version of the same problem in my job: I publish documents using Acrobat Pro, but most of the end users have only the free Reader. When a user tells me that docs are not working (i.e., printing, displaying, navigating, etc.) correctly, I can't simply say, "works fine for me." I have to figure out why a document that works in my "playground" doesn't work for a Reader user on a system with potentially different settings and connected to a different type of printer. And if lots of users reported lots of different problems with our documents that we couldn't diagnose and fix (thankfully that hasn't been the case), we'd have to change our tools... because having things work for the end user is the only thing that matters.
I had a friend in high school who used to like to mess with people at public Information booths by asking funny questions, and one of his favorites was "How many Frenchmen can't be wrong." My question to the SixApart and/or Scienceblogs support personnel is How many Pharyngula readers can't be wrong?
The last time registration was required, I had terrible problems. Like some others I couldn't get Movable Type to send me the confirmation email.
So I went with OpenID, which at least worked spasmodically - it usually took several tries to sign in each time.
Since the registration was turned back on this time, Open ID has worked like a charm.
I just wish someone hadn't beaten me to my favourite username before I signed up.
Oh and the problem Steven Mading just wrote. I have gotten that one too. Believe me, I wouldn't deny it if I just forgot my password.
#69:
Again: Absolute nonsense! (See #47).
If you're a programmer and you have that attitude (giving up when you can't reproduce the problem), then I feel sorry for the people you're supposed to help!
I accept "can't reproduce the problem" as an excuse for the fix to take longer than expected. But that's about it. This is where you become creative. This is where experience and gut feeling play a role, this is what separates the REAL programmers from the code monkeys.
Well, I just went through registration. It actually worked for me, which is a bit of a surprise (running Google Chrome 4 on Linux, BTW.) Frankly, I'm surprised that the registration keeps spammers out; it didn't have any kind of verification to make sure the applicant was actually human. Most will have at least a basic CAPTCHA, and even those have a hard time with many bots (case in point - the phpBB captcha system.) Maybe the only reason the registration keeps bots out is because it doesn't work most of the time...
@71
Yes, I received the same bounced email message. I usually send the already overwhelmed PZ the same info I try to send to the webmaster at scienceblogs.com because I know it won't go through to the webmaster. Apparently PZ can forward the message, so the exact record of what happened (or failed to happen) when I tried to sign in does arrive in the inbox of some hapless coder.
You guys saying that if you can't reproduce a problem you can't fix it are only proving the point that you have to work to reproduce or track down the problems the users experience.
I just use my google account.
CAPTCHA CAPTCHA CAPTCHA CAPTCHA
CAPTCHA CAPTCHA CAPTCHA CAPTCHA
CAPTCHA CAPTCHA CAPTCHA CAPTCHA
CAPTCHA CAPTCHA CAPTCHA CAPTCHA
CAPTCHA CAPTCHA CAPTCHA CAPTCHA
CAPTCHA CAPTCHA CAPTCHA CAPTCHA
CAPTCHA CAPTCHA CAPTCHA CAPTCHA
CAPTCHA CAPTCHA CAPTCHA CAPTCHA
CAPTCHA CAPTCHA CAPTCHA CAPTCHA
"ctrl v" several more time.
Really, is that hard to add a captcha system on here.
John Harshman,
The Movable Type password reset email used to have the subject: Password Recovery and it comes from the same sender as the confirmation email (system@scienceblogs.com) if you want to look for that in your junk/spam folder or add it to your list of safe email senders.
Indeed, the username you enter to sign in should be the one you originally registered with (Movable Type now differentiates between your modifiable screen name and permanent username).
However, the password reset/recovery tool only asks for your email address, so I would think the only possibilities are:
--the scienceblogs.com email is being caught by your inbox spam filter (which you can get around by adding "system@scienceblogs.com" to your safe contact list),
--scienceblogs.com itself is not sending the email as it says it is (you're screwed), or
--the scienceblogs.com is blacklisted and any emails they send are being automatically deleted by your email service provider (you're screwed).
I'm still shocked that Six Apart came on here a few weeks ago and tinkered with things but did not address the problem of missing Movable Type confirmation and password recovery emails, the mangled OpenID names, or the Garbled Wall of Silence for TypePad users even after being told about those problems repeatedly.
Someone willing to actually make a phone call to scienceblogs to see if you can get them into action?
I'm off to sleep. (5am) boy oh boy won't tomorrow (4 hours away) be a big ball of fun.
Seems to work well with google (mac, os 10.sth), fwiw.
On a different angle: it's nice to see you (PZ) keep the comments as open as you are, even though thread maintenance must eat quite a bit of your time. An informative resource and a testament* to your point of view. Thanks.
(*I'll leave it to you to decide what kind, "new" and "old" being taken.)
<headdesk>
<headdesk>
<headdesk>
<headdesk>
<headdesk>
@ necronomikron #83:
And it shows! Which is no reflection on you, but instead on a newly introduced system problem - that of the missing logos for the extra sign-in options. The google image is one of those which is missing from its assigned location.
Meanwhile, the way that the typepad/TypeKey logo keeps swapping (within the same comment headers on the same comment thread on different refreshes of the page), between a blue and white key+keyhole and an anonymous black bod on white, shows how the Sb system keeps forgetting that typepad/Typekey is something it's supposed to know about.
I still suspect that Sb has more than one server with more than one copy of the software on it (with the versions failing to be kept in line) and only some people get the most recent/valid version of links and hence can get in. Dodgy caches can exist anywhere between what the developers imagine themselves to be outputting and what the user actually receives.
This has certainly been the cause of some of the BBC's intermittent problems. Those of us in the know could bypass some of the issues by explicitly addressing the server paths which were found to be working instead of accepting whatever we were randomly issued each time.
Re: Movable Type emails
I even got one of the guys from the Secure Gateway team here at work to see if our firewall was eating the emails - nope. And he tried to register with Movable Type and couldn't get it to work either :-)
neon-elf,
Thanks. That pretty much narrows it down to scienceblogs.com itself being responsible for not sending out the "missing" confirmation emails, doesn't it? Nothing can be done about it from the user end if that is the way things are.
Well, then, I'll comment online when I can, and when I can't, I'll be commenting by ESP. To my dog.
Sign up via Movable Type worked for me, no issues. I'm running Firefox with NoScript.
This is probably a dumb suggestion (I don't know jack about this stuff), but could we move the commenting off Science Blogs by providing a link after each post to an external, well-functioning forum elsewhere?
me, I hate typing oUt cRaZi Captchas, and now that I've bothered to register it's fairly easy.
I just registered and it worked fine.
Look on the bright side. If you've got an incomprehensible name, your arguments will be judged solely on their quality, and not on your past mistakes.
The fact there is no registration link, only a tiny Sign In above the commentbox link may hinder ppl trying to register.
Typepad registration proceeded without problems, and allowed me to specify my nick.
Of course, personal experience is no proof this actually works.
charley | December 15, 2009 2:15 PM:
There are quite a few blogs that seem to happy with that solution. However, I find it to be such a nuisance that I am seldom a repeat visitor at such blogs.
This is all a bit of a conundrum. I come to this site principally for the loons (Yay Jesus!!!) and for the head explosions their nonsense causes.
Use a Yahoo! account to get an OpenID address. Sign in with that. Registration problem solved.
It's unfair to talk of programmers this way, most (the ones who I know anyway) use alcohol as their drug of choice ;)
Among other things, you'd think the authentication server would be logging failure errors. Sometimes error codes can be helpful.
Also, I have had much better success with Firefox than with IE. Word to the wise.
For the longest time I was able to sign in and comment using TypePad. Then, yesterday I was continually getting an error whenever I attempted to comment.
Followed by a page full of broken code. The strange thing is, if I physically went to the Typepad.com page I was logged in there. But if I tried to comment here and selected Typepad I received errors.
In frustration I created a Moveable Type account. It has worked so far.
MacOSX 10.5.8, Firefox 3.5.5
Why don't you just hire a JUNIOR developer with half a brain, for a fraction of the cost, that could solve your registration woes in a single night? Or find a particularly dedicated individual willing to volunteer their time?
Hey, I figured out how to register. How had could it be?
The simple fact that my openid name appears as "krc [clowersnet.net]" and others appear as simple url strings is evidence enough that SixApart is incompetent. Every other place where I use openid correctly gives my name as "Kelly Clowers".
I really have to laugh at the people saying "just do X and your problems will be gone". I have had little trouble commenting, but it is obvious that the problem is highly intermittent. Given how long this has gone on, I think SixApart just needs to junk the current code and write from scratch something that is clean and maintainable (or get someone else to implement it, since I am not sure SixApart is actually capable of that).
Yep. I ran in that one many, many weeks ago. I gave up trying to inform them I cannot access the
Living the Scientific Life (Scientist, Interrupted) blog. I keep getting:
I've been getting that for months now, and Teh SciBorg doesn't want to know. I suspect grrlscientist might like to know, but since I can't read her blog, it's a bit hard to locate an e-mail address.
A suggestion I've made several times is Teh SciBorg's IT staff should have a blog or similar. That provides a two-way channel for constructive communication, rather than the current situation of trying to tell them of problems or make suggestions, and either being silently ignored, getting yet more errors, or both.
I tried it using a google id and now I'm trying it using a yahoo id.
FWIW, my recent inability log in via TypePad was resolved by logging out from TypePad.com and then signing in here again.
Speaking of sign in, a few people have reported the "sign in" link not showing on their system. One could see it with Firefox at home but not with Internet Explorer at work.
Yep. Looks solved to me.
Another fail.
Can you rig little red and green flags at the top of the frontpage to indicate when nonregistrants can and can't troll spam blather comment?
Have to admit that display of those unwieldy URL's rather than user names of people coming through google and yahoo is getting annoying.
Any chance you folks could sign your posts with your handle until SixApart fixes this? (if they ever fix it, of course)
Agree that a CAPTCHA mechanism would be the best way to solve PZ's spam problem. Registration puts up that little bit more of a barrier to trolls, and they are just *so* much fun to play with around here ;-)
No trouble registering with moveable type.
Trouble with reading pharyngula with most browser types/script settings.
1. Two of my browsers crash reading with pharyngula. If I turn off scripting and activeX, whatever those are, then it works. Except lots of other websites don't work.
2. One browser reads fine. When I hit comment, get an "invalid request" message. This didn't happen until they fixed the registration problem.
Many of the sciblog problems are due to not being compatible with a lot of browsers and settings, particularly the older ones.
OF ALL THE MANY WEBSITES I VISIT, SCIBLOGS IS THE ONLY ONE WHERE ALL THIS WEIRDNESS HAPPENS.
This is a problem with lack of good coding unique to sciblogs. As to what can be done, well, I'm just a lone voice calling out from the wilderness. There are undoubtedly many others who would except......THEY CAN'T ACCESS PHARYNGULA FROM THEIR COMPUTERS. Well, duh.
Ok, login via Google looks to be working ... for some small value of working. No matter, I can live with that.
So let's see how exactly the result looks like ...
Works as expected, except ridiculously slow.
So I just have to remember to sign off.
khms
For a long time I was able to sign in using TypePad. Then just recently it told me that I needed to give a valid email. Several people, especially Nerd of Redhead, gave a fix. I wasn't able to implement that fix because I was NEVER able to find the switch to toggle.
For a long time I wasn't able to use Movable Type because the confirmation email NEVER showed up. I looked in all the folders (incoming, outgoing, spam, drafts and trash) and it was NEVER there. For some reason when I gave my gmail address I got the confirmation email but two other email addresses couldn't get it.
Hey, all you computer weenies who whine "your 2 stoopid 2 do it rite," fuck you. I'm a reasonably knowledgeable user and I can't use the fucking software. That's YOUR problem, not mine. You're supposed to give software that works reliably for us iggerant users, not shit that needs a Masters in Computer Science, twenty years experience, and sacrificing a virgin gerbil to get to run in a half-assed manner.
blf, I left a message on grrlscientist's blog pointing her at your comment. Her contact is GrrlScientist [at] gmail [dot] com.
I'm not sure if the concept of partial reinforcement will work here, but I say let's conduct an experiment. ;)
Oh really? I bet you forgot to sacrifice the chicken....again!!! LOL, (joke)
I'm probably an above average lay user, and sciblog almost defeated me. I tried about 20 fixes. One killed my computer OS dead. Very dead. I spent hours reloading software to revive it.
After that experience, not going for extreme measures anymore. PS: I haven't yet summoned the nerve to ask my computer whether it saw the blinding white light and met its forerunners, MS DOS and Apple OS.
The point is valid. One shouldn't have to be a hacker taking a break from identity theft and reading the Pentagon email to cruise the nets and read pharyngula.
I couldn't get an OpenID URL to work, but my Google name works. So much for anonymity on teh innerwebz.
Hmm, Google-Account seems to work. Would be great if the username wouldn't look so stupid, though...
So that's why I'm having trouble. I'm stuck with a Bachelors in Computer Science, didn't specialise enough to figure it out ;)More than anything I've found the software unreliable. It's not that it works if you do the right thing, but that it sometimes fails when doing the same thing. Clearly the problem is not a PEBKAC issue, but a software issue.
Raven, there is something very wrong if a supposed fix for a website registration issues destroyed your OS. Was it something someone here said to try? Or from somewhere else? Or was it just something you thought of and tried?
Anyway, that an your statement that you have to turn off ActiveX and JS to not crash on scienceblogs, makes me think that you have some sort malware infection, or else your computer is just really messed up. You should probably get a good tech to check it out.
It was just me going through my bag of tricks and making new ones up.
It's possibly malware although I sweep it periodically and didn't find any.
And it could be messed up. This is an old but high end custom made system (not mass produced) that is fast on the nets. But it has been heavily used and added to and is approaching obsolescence. By now it would be more cost effective to buy another box.
I'm really not going to noodle around with it too much anymore. One of the these days it will just die for good. Computers just get old and obsolete like everything.
Well, got LiveJournal to work and now TypeKey works. Still having issues with Movable Type. The email never shows. What is the email address fix Nerd or 'Tis??
Was able to resister/login with my Google account without too much issue. Lets see if I can post...
TRIED to register last week - Still waiting for that "confirmation email." :/
I was also able to login with my Google account :)
Here we are, third or fourth try.
If you didn't get it pretty much straight away, it's not coming - ever. That's now another dead login name which neither you nor anyone else will be able to use (because it's already been assigned as far as the system is concerned but is stuck in unconfirmed status). The would-be system fixers have never issued (via PZ) any mention of clearing up old messes.
Incredible but true: less than ten hours after registration is turned off, Dennis Markuze infests the Oral Roberts thread (from comment 67 onwards).
And he refuses to give the Right Rev. BigDumbChimp, KoT, OM, his FINISHED!!!
Perhaps Science Blogs should consider hiring professional programmers. I've been writing software for critical applications (steel mills, nuclear reactors, medical equipment) for 25 years. I've never once uttered the words "I can't reproduce it therefore I can't fix it". If I did, it would probably be my last day as a professional programmer. Lots of problems can't be reproduced in a controlled environment. Get used to it. Add code to capture user input. Record the name of the browser and OS in a log file. Log every time someone tries to create and account and fails. There are lots of options. "We can't fix it" is a cry for children. If you're getting paid to program, then fixing the bugs is part of what you do. If you don't know basic debugging techniques (like instrumenting your code), then it is simply not reasonable for you to expect to be paid for the code that you produce.
I hope this comes through. Grrr. Problems.
Yeah. Due to the phenomenally arcane magic used in the top- and sidebar ads, it basically stopped working in IE7 shortly before IE8 came out. (It does work in IE8, though. It even works in Safari for Mac OS 10.3, probably because that browser doesn't even try to parse the magic and just ignores it, or something... ~:-| )
You are not logged in.
I had tried to sign up with Movable Type using gmx and lycos. Never got a confirmation email. Last week when TypePad insisted I hadn't given an email address (which I had), I used gmail to sign up with Movable Type. That time I got a confirmation email.
Checking just arrived is the door held open yet?
If it is...Like you say if someone had a comment and lurked for an opening, that would suffice.
TypePad seems to have screwed up its settings recently; if you have an existing account, you need to go to your account, check the "Share your email address" box, and save your changes (I too got caught out briefly) to be able to sign in here.
Other than that, I've found it reliable.
It did come through, but only because registration is switched off – you're not logged in.
Oh boy! If this works, thank you to everyone who made me laugh so hard with GOATS ON FIRE and other recent hilarity. I'd been feeling a bit gloomed and you dispelled it.
@133, nobody said the problem wasn't fixable, just that if it wasn't reproducible at a local level makes it harder to fix.
We found several issues with the site itself after 30 minutes of looking suggests the admins in charge of this site are incompetent.
Might I suggest pitching reCAPTCHA to SixApart?
http://recaptcha.net/
It's free and helps convert old books into electronic format.
Would it be feasible to give admin powers to trusted commenters? Say the recipients of the Order of Molly are given the power to delete offending powers. That should allow registration-free commenting while increasing the amount of people watching threads by about 30 odd.
Just a thought.
Bah, that should say "power to delete offending posts". Sorry.
Testing to see if I can comment. If you can see this, then I can!
I guess you just have to keep trying and trying until the server finally relents and accepts the sign-in submission. Pharyngula is worth it - but ugh to those programmers. If mine they'd be on measured mile.
Tried registering again and finally received a confirmation e-mail, after many failed tries over the last few months. Yay!
Well, OpenID authentication succeeds, for me at this time, but is still broken: the identity given on this comment is *not* my OpenID. It is a transitory identifier for a single transaction, that Moveable Type has mistakenly used as a persistent identity.
I'm glad that OpenID is a consideration, and an option for commenting. I recommend anyone having trouble to try their OpenID (they probably already have at least one that will work). Just don't expect the resulting identifier on this blog to make any sense: it might, but it might not.
Next year you can just replace the commenting system with Google Wave. :)
I suspect raven's PC crashes might be caused by lack of memory/ processing power, if it works when all the memory-hunbgry gadgets are turned off. At work on an P4 with 1 GB ram and IE8 I have to be very careful which threads I load, anything over 200 comments makes the fan go off and attracts attention...:-)
SciB stuffs way too much shit into the sidebar menus, and I suspect some older PCs just fold when trying to load it.
And if this problem turns out to be located at the interface between SciB and Sixapart, it's never going to get solved, can't even talk to one party, let alone get them together !
as a fellow blog writer, i will add my own voice to the chorus: i am OUTRAGED to see that scienceblogs has the audacity to ignore our readers when they report problems, such as being unable to view a blog! blf is a long-time devoted reader of mine, so his complaints are even more worrying since i doubt that a new reader or a casual viewer would bother to complain about being unable to view my blog. this lack of concern for the very public that they are supposed to provide information to goes against the (presumed) core values of the organization itself.
It doesnt make sense that blf can't access one particular blog only, unless he/she is attempting directory browsing or something.
HTTP error 403- Forbidden
Have you cleared cookies, tried different browsers etc ?
Yeah but don't you know, the 1 IT guy they had, well, left.Maybe they haven't got around to replacing him/her yet.
In my long career as a developer I believe I have fixed COUNTLESS problems that I was not able to reproduce. There fixed it for you. Your faith based development does not impress me.
If you can't reproduce the error, and it occurs intermittently you cannot know whether or not you fixed it. Software development is no different from other sciences. It is not Black Magic, even tho' some people would have you believe differently.
I never said it wasn't. I said that if you want to address the issue, you're going to have to be a lot more constructive than "It doesn't work for me / it works fine here". Whining about it on Pharyngula isn't helpful for resolution. MT now offers 9 different ways of signing in. Chances are one of them will work. I agree that using the OpenID url as author name is a bit stupid (just leave me a box I can type FlameDuck in once I've authenticated with OpenID, kthnx).
Actually I think it's more likely that their email server must have some kind of problem. Most MX servers have some pretty strict rules for what kind of mail they accept (to combat spam, obviously) and MT might be using old software that isn't completely RFC compliant.
Not if the user in question doesn't tell you what shitty browser he's using! Also Windows doesn't normally allow you to have more than a single version of IE installed at any given time, so testing all the possible IE configurations is impossible. If you're not going to provide information about what browser you're using, what "security" settings you have and have not enabled and what strange 3rd party "plug-ins" you've chosen to install (or that IE has installed behind your back, without asking you), there is no way a developer can duplicate "the entire range of user experience".
It's what separates working systems, from nightmare ones. If you have to rely on divine inspiration to fix a problem, you are making it worse. Do you work at Siz Apart? Because the kind of mentality you're expressing is, in my experience, the kind of mentality that inevitably leads to systems expressing non-deterministic behaviour.
I doubt any of "us" are saying any differently.
And I do. Unfortunately for you, I don't work at Six Apart. If you live in Denmark, and want a good and easy to use on-line mobile phone company on the other hand, I could probably hook you up. :o>
@ Rorschach #152:
I was going to post about the 403 Forbidden error last night - but I was just so very very cold and unenthusiastic.
ScienceBlogs seems to return the 403 Forbidden error randomly, ie non-deterministically - from the user's point of view. It certainly isn't simply about directory browsing because, on the whole, Sb lets you do that!
In particular, 403 Forbidden is what you get/got a lot when trying to login with MovableType. It would happen at the very start of the process - long before getting a chance to enter a name, email and password. So it's a separate bug (at least at the user end) to the one where the confirmation email goes missing. It makes it look like the user is completely banned from using MT at all. And it can go away again for the same person operating the same browser with the same settings on the same PC. It's definitely not the user's fault.
I think it's again because the various pages are intermittently being served up with faulty URLs embedded in them - including that initial sign-up one. We even know what those faulty URLs are in some cases (ie when the page is presented to a competent user who checks what's hiding there)! But no Sb/Six-Apart code-monkeys have ever seemed to care. What we can't so easily know is what's on dodgy pages which Sb and the SixApart sites serve to themselves and to each other behind the scenes during the sign-on process and which then cause a cascade of different errors to appear.
scienceblogs.com/mt-comments.fcgi?__mode= ... is always bad
scienceblogs.com/mt/comments?__mode= ... sometimes works (depending on what follows)
Note that, for this issue, it doesn't seem to matter much which browser you're using. I have several here and have tried more than one on pharyngula with matching results. "It's your browser" is largely another red herring where Sb/SixApart (or passing trolls) try to blame users rather than their own code.
I suspect the code is written in some rubbish modern language which assembles its pages and their URLs in pieces from tiny fragments in an obscure way. Otherwise it would be "easy" for a software engineer to root out all instances of the faulty (old?) URL across the collection of files and server copies of them - regardless of not ever getting served dodgy pages themselves or seeing the error from their privileged position.
"It's your browser" is largely another red herring where Sb/SixApart (or passing trolls) try to blame users rather than their own code.For the record let me clarify that I haven't ever said "It's your browser" has anything at all to do with the issues being encountered. It was used as an example. A 403 error is however a http return code, so it could be because your browser is failing to authenticate you properly. It could also be a web service call that doesn't fail elegantly. Without knowing more about Six Aparts software or server setup it's impossible to tell. However 403 errors are easy to reproduce, because they can only ever possibly occur in very specific circumstances. Where I to hazard a guess I would suppose that when they reorganized their servers, sometime between May and July of this year, someone forgot to update the mod_access configuration in the httpd.conf or .htaccess on one or more of the Apache servers running the site.
This would explain why it happens intermittently (because whether you get it or not would depend on which Apache server the load balancer forwards you to) and would explain why they can't reproduce it in their developer setup (because the configuration files there, would be correct).
Odd. I was about to suggest that it's probably because it's written in some archaic language like C++ or perl, which still has to rely on FastCGI to provide web applications. In fact, what do you know, the wikipedia page for Moveable Type confirms that the software is written in Perl and PHP.
In either case, the 403 forbidden error is caused by the httpd server (in this case Apache), and is not caused by the programmers, or their poor choice of programming language.
'Tis Himself, hi,
The problem/solution is exactly as John Morales describes:
Once the account is in a state where you are not sharing your e-mail address, you can't log on via ScienceBlogs. You have to go directly to your TypePad account and fix it.
Just as I said earlier.
I regard Perl and PHP as the trashy modern stuff though! C is better than C++ etc.
The errors from the two versions of the mt directory construction has to be caused by coding. It can't be just a failure of servers to communicate at all.
It's sciblogs fault. Pure and simple.
I visit a dozen or so websites every day. Some commercial, some not. For many reasons.
The only one with multiple, reoccurring, and endlessly mutating problems is.....SCIBLOGS.
One site that I've frequented for many years has a simple format, requires sign in to post, and has never, ever given me any problems. The owners know what they are doing and do it well. It has changed a lot over the years but everything always works.
It is possible and common to run a website that works. Hopefully, someday SCIBLOGS will wake up and realize this.
I have just managed to reset my password on TypePad -- I actually received the email from them -- and was able to sign in. See how long that lasts.
By the way, when I previously went directly to the TypePad site (not through Pharyngula), it was glad to accept my old password, though it wouldn't when going through Pharyngula. Go figure.
Is it possible to sign in with OpenID but to have one's nickname appear instead of one's OpenID identifier? I have a nickname set with my OpenID provider, and it should provide that information to Sci Blogs as I sign in. Failing that, Sci Blogs should be able to associate a nickname with my OpenID identifier.
(My normal identifier isn't too bad: http://timothy.green.name. But I'm having problems with it at the moment, and am temporarily using Google as my OpenID provider, and that produces painfully ugly identifier URLs.)
TRiG.
Testing.
> the identity given on this comment is *not* my OpenID. It is a transitory identifier for a single transaction, that Moveable Type has mistakenly used as a persistent identity.
In the meantime this has evidently changed. Moveable Type is now showing my actual OpenID for my identifier. This is an improvement; thank you.
Better would be for Moveable Type to use the OpenID protocol for getting the user's requested display name, as others here have requested and as other sites do without trouble. One step at a time, I guess.