"Why this isn't fixed yet."

Zoe and others are agitating about bugs. While I made progress today, it's not always clear why some bugs get worked on and others don't. With 100 active bugs, and another 50 in discussion or reopened, I'm making a variety of choices. Most of these choices are boring and trivial, like not fixing it because I don't think it's as important as other bugs, or, if as important, much harder to fix.

Anyway, I'm going to go ahead and add a brief note to every bug about why it's not yet fixed. Please don't be offended by the offhand tone. My goal is to fix every bug. But if you want me to reveal my mental process in the order I fix things, I will.

Thanks. I don't generally care what order you work on them, with one huge exception: data loss issues should be a top priority and should be fixed as soon as possible. As in, don't work on anything else until the problem is solved and the lost data is recovered. I didn't even lose any data with the recent multiple-reading-date bug, but it's the principle of the thing. I want to know that my data is safe.

I was also wondering whether it would be helpful to add some way for users to mark bug priority. Say, each user can specify whether a given bug is low-priority, medium-priority, or high-priority. These counts would be cumulative, so you could see that one bug was marked low-priority by 8 users and high-priority by 2, while another bug was marked low-priority by 2, medium-priority by 3, and high-priority by 5.

For the "assigned to another employee" bugs, it would be nice if those people could stop by and give a real explanation ;)

I'm thinking of changing my name to _Silent_

Thank you thank you thank you.

I completely understand that some bugs get higher priority, some bugs take longer to fix, etc. What "agitates" me is when a bug sits with no response or explanation from any admin. So yes, I hope this will be better.

Though seeing all the responses you've put in all these open bugs, it's obvious how much time you can spend saying you don't have enough time to fix it.

I'm glad that you are leaving notes on a lot of the bug reports, but it's starting to get frustrating. What good does it do for us users when all you put is "CH's" ? What the heck does that mean? I *think* it means that CH is the one supposed to fix/handle it... But does that mean you've talked to CH about it, that CH needs to response, that CH is unavailable.... What does "CH's" mean?

If you are going through all this trouble to make notes in all these bug threads, because of us users who get "agitated" about it, it seems kind of counterproductive to put one-word notes that users can't understand.

I trust Tim et al to prioritise sensibly. I'd rather time was spent fixing stuff than publicly commenting on it. Just sayin'...

Apologies for "CH." Yes, it means CH is responsible for the bug (see the assignment section on the top-right). I shall encourage LT employees to put their own notices in as soon as they can.

>8 mart1n:

It was an exercise for me too.

Incidentally, I need to finish up by adding the "reopened" and "discussion." I fell apart there at the end—had to nap before getting up, now, to take train to NYC. The others will have to wait a bit.

#8
I'd rather time was spent fixing stuff than publicly commenting on it.

Of two bugs I am particularly interested in, one I reported a couple of days ago and the other was reported on the 28th September. I'm very glad to have got a response in each case. Thank you, Tim!

(And I think that one public comment after three months is hardly wasting time).

>7 Heather19:

CH is Chris Holland.

Also ...

Thank-you, thank-you, thank-you Tim.

I'd love to be able to give a weighting to bugs, but I don't think it's realistic to expect that. Also, I agree with Zoe that data loss should be priority number one.

Go out for dinner.....

No! Stay at your post! We'll have a whip-round for a pizza. ;-)

I'll add to the choir of "Thank-you, thank-you, thank-you Tim." I'll even add to a whip-round for a dinner :-)

(BTW the use of CH made me wonder if Tim CHown's these bugs to Chris :-)

15>

BTW the use of CH made me wonder if Tim CHown's these bugs to Chris

Groan.

Thanks for all the time you put into these problems.
I was surprised to see an answere on some messages, because they weren't of that much importance. Mostly I just wait untill youv'e the time to fix it, because I can't see from here which prblems there are.

So, a week later... I'm really disgusted that you still haven't done anything about restoring lost data, well over a month after the fact, while saying that it's "crazy" for book shipments to take more than a month in the face of severe winter storms. At least BD can say that the books are now out of their hands; what's your excuse?

You told us that you cared about our data. You lied.

I get where you're coming from Zoe, but I don't actually see people here complaining about it. It seems to me this was restricted to people who already fixed it. You, for example, didn't have the problem.

I'll fix it tonight and we'll see how many people use the tool. My guess is less than five, to see how it works.

See this message; there's still something wrong there. Also, some people may not even have realized that data was lost.

More generally, how are we supposed to trust you with our data in the future? It doesn't matter if only five people lost data, or even if it was only one person. Unless you treat all data loss seriously and immediately, we all have to worry that next time we might be the unlucky ones who lose data that you don't care enough to recover.

I really don't think you want to go the route of "this bug didn't get enough complaining". That could get ugly fast.

I never really complained about it because I thought Zoe was doing a good job expressing the position.

Zoe, I think that's a very good point. I don't know if I've "lost" any data with the multiple-read bug, but at the moment I cannot retrieve any of it. I am hoping it's in the database, but thanks to these bugs, not accessible currently. So while this bug is, I hope, not destroying data, it's certainly making it "lost" to me until it's fixed.

Zoe makes a very, very good point. I didn't lose any data with that bug so I haven't been complaining, but I kinda assumed that Zoe's constant reminding would be enough to spur action. I guess that's a wrong assumption. It doesn't matter how many people are complaining, what matters is the simple fact that if data-loss isn't treated seriously and quickly, people in general will start to lose trust in the website. No one wants that.

Re 21: Seconded. I'm just tagging things with fixme and moving on. When it gets fixed, it gets fixed.

>24 theapparatus: The "fix" has been partially implemented already--if you go to the edit page of each affected book, there should be an option to show all the reading dates they ever had. But I'm still hoping for something more general, so that people don't have to individually inspect each of their books to determine whether any data has been lost.

Though this is an aspect of LT I don't use myself, I think any data-loss is very bad.

Tim> You forgot to comment on why this isn't fixed yet:

http://www.librarything.com/topic/102922

Bug in a major component of LT, open for three months, nary a peep out of LT staff.

Kinda frustrating.

This was reported elsewhere. I've assigned it to Casey. He told me he'd fix it as soon as he could. I'll follow up on Monday.

This was reported elsewhere.

Err.. where? You mean the report I linked to or stuck in some other random thread? By reported you mean someone at LT responded, or just that the problem has been reported to you? Which doesn't make sense, because I linked to the report from three months ago... so, I guess I just don't understand what that's supposed to mean.

My catalogue still suffers from data loss despite the attempt at fixing it.
I tried to use the tool that was supposed to help show the dates that had gone missing on one of my books which suffered from data loss back when this was first reported.
And at first the dates I expected showed up. I then started reentering these deleted dates, but after the first edit, the tool no longer shows the right dates. All the dates are now the one I've just entered.

This is not a fix! This is an outrage.

The work in question: http://www.librarything.com/work/15386/edit/23244938

I even reported this problem over a month ago on the original thread, but no-one has reacted.

Despite all the attacks, I want to emphasize that I do believe LT's abandonment of the user-data-protection principle is the worst thing that's ever happened to the site. I know it's easy to let it slide in "minor" cases, but you can't really hold principles halfway. Either all user data is sacred and preserving it is your top priority, or it's not.

Or Tim could let us know ahead of Tim which data is sacrosanct and which is less important.

I will take a look at this topic again, and improve the tool. I don't regard a shortcoming of a recovery tool built to help a small number of users--that it allows you to see all prior data on a book, as requested, but only until you replace it--as an abandonment of user data.

Thanks. It's really a matter of response time more than anything else. In cases where data loss is involved, I'd like to see you respond immediately, even if just to defend the existing recovery tool as sufficient. Leaving these sorts of bugs open for months at a time without even a note looks very bad, and raises all sorts of doubts about how similar situations will be handled in the future.

No, I hear you.

Thanks.

For future reference, I'm also concerned about hidden data loss. I don't want to have to pay attention to Bug Collectors to see whether I may have lost data when I wasn't looking; if there was a bug that may have affected my data and that requires action on my part (like using this recovery tool), I'd like to be notified about it via profile comment with instructions on how to proceed.

>36 _Zoe_: Absolutely right. With 4000-odd entries, all sorts of corruption could sneak in un-noticed.

@35, 36, 37

A solution strategy would be to spread the risk of data loss.

Why not make it possible for individuals to take the extra step of creating a here-and-now copy of the state their collection. A button on the webpage can save a file to the member's Download folder that they can keep safe with their own solution, like Apple's time machine. To be fair, expectations ought to be mindful of LT's funding model and the effects of Murphy's Law on data.

Why not make it possible for individuals to take the extra step of creating a here-and-now copy of the state their collection.

This has been possible for many years -- look at Import/Export under "More". The tab-delimited export is far more complete than the CSV version.

@38 We already have that although not in button form:

http://www.librarything.com/more/import

I tried the tab-delimited export and it wants to save as an xls file (Excel). However, when I try to open it (in Gnumeric and OpenOffice, on Ubuntu) it says "non-supported file format".

41>

It's tab-delimited, not XLS, despite the name. Try saving it as a .txt and see if OpenOffice can open it that way.

LibreOffice/OpenOffice Calc on WinXP and Win7 opens it fine. I don't know why...well, it's different because Ubuntu doesn't make the assumption about .xls files, I guess.

I changed the extension to txt when saving and it looks great :D

@39, 40

The import & export feature when combined with a spot the difference among files mechanism should allay fears @36, 37. The view history webpage on wikipedia does this sort of comparison.

@45: When you want to view differences between two (text) files on Linux, you can just type "diff (filename 1) (filename 2)". I don't know if there's something like that for Windows though.

@46: On Windows, access to diff(1) is reachable via Cygwin or Windows Services for UNIX.

On the Mac, there is diff(1) and Developer:Applications:Utilities:FileMerger.app. On Emacs, there is ediff.

To find text in apparently non-text files(*), od(1) is a place to start, at the Terminal.app on the Mac or Linux console, run the command `man od' to read the manual. Listed in the `see also' section are two more related commands.

* e.g. corrupt data

I don't know how to post a new bug, and I don't want to bug Tim directly, it is disappointing to find missing images, like

http://www.librarything.com/userpics/SamuelJohnsonLibrary-big.jpg

Perhaps the following tools could be used to store pictures (book covers, too) more permanently:

http://www.dspace.org/
http://duraspace.org/
http://islandora.ca/about

48> Is this what you're looking for? What link did you follow to get to the URL in your post? LT stores their pictures pretty well, from what I've seen.

@49 thanks for satisfying my curiosity, I wasn't looking for anything, just following missing links

That busted link I found in one of the blogs dating back to mid-08.

Another missing link is

http://www.librarything.com/pics/blog/tagging1.mov

a univocal set of URLs to these resources would be nice to have

Another example of a missing image is above the paragraph mentioning Madonna's Sex on page

http://www.librarything.com/blogs/thingology/2008/08/

I get the tagging movie just fine, and I don't see what's missing in the blog post about the Bookfinder report - there's no picture intended to be there, as far as I can see.

I suspect - I didn't go look - that the blog post was about the early setup of the current image system, or possibly the system before that, so the link was correct at the time but hasn't been updated to the way things work now. I found the picture (probably the picture intended) by going to the Legacy Library page and clicking on the picture displayed there, which would be the normal way to find the pic. I suppose they could go back through all the blogs and update the links so that they would still work...but I'd rather the programmers spent time working on new features and bug-squashing rather than updating old links.

Sorry. Doesn't rate as a bug, or even a real problem, on my radar.

> Sorry. Doesn't rate as a bug, or even a real problem, on my radar.

The tagging movie is viewable, now. The broken image links are spots to get used to and ignore, ok.

Broken image links in blog entries from three years ago - yeah, they can be ignored. In my opinion, anyway. Broken links on active-use pages - that's a problem and should be reported. You see where I'm coming from?

Hi! Just did just run again into an older problem:
"Bug Collectors: normalization: ignore « leading / trailing spaces » « spaces / whitespace » in « Common Knowle"
"Recommend Site Improvements: CK cleanup regarding leading space(s), training space(s), multiple spaces"

@53 what's an acceptable excuse for seemingly random broken image links to book covers, for example? if no-one sees where that is coming from what else could be disappearing and ignored? i don't want to get into an argument, l see avoidable trouble

@54 a regular expression should do that for you

>56 vy0123: To my understanding the datas are at the server site; in th LT (CK) database. Whoever has access there might run the relevant script. I never managed to learn regEx. What is the replace command for cleaning a) whitespace header, b) multiple consecutive whitespace occurences insite other relevant text, c) whitespace trailer?
Both examples for JavaScript and for mySql might be helpfull.

55> Have you seen missing book cover images? All your examples were from 3-year-old blog posts.

ETA - There have been missing cover image bugs; all (that I'm aware of) that have been reported in Bug Collectors have been fixed. If it's showing up again, it should be reported again so that it can be dealt with.

《57 look it up in a book or search engine, test and verify

@58 yes, there are missing book cover images, e.g. zen and the art of motorcycle maintenance, that one has a missing preview in the tiny box set area, perhaps the growing pains of '08, '09 lost more than usual, I find sometimes if I truncate the URL or click the search button on the missing error page, I can find the target

60>

That sounds like this bug.

61<

Sounds like that bug but looks different. The one-pixel cover image is unlike a missing link represented by question icon. Some things to try on the LT system:

• after a cover image is uploaded, flush that file to disk
  
• pair a checksum with cover image
  
• routine fscking
  
• maybe use a product like this®