Este tema está marcado actualmente como "inactivo"—el último mensaje es de hace más de 90 días. Puedes reactivarlo escribiendo una respuesta.
1pivox
I will start this thread more as a test. The idea is, whenever you come across a post by Tim and it does hold some relevant data (which it almost always does), you add a link to that post here. The objective being to have a centralized place with all of Tim's posting, so you won't miss any of his comments on any of the relevant subjects (ideally, this could be done somehow dynamically - a group that automatically adds a link to any post by Tim whenever he posts one).
Ah, and pleas don't comment his posts here. Go to where the link leads you and comment there.
And If you want to leave a comment about this thread do it in about "Really stalking Tim" rather than here.
So here it starts with the one of Tim's post I stumbled upon today
____________________________________________________________
Tuesday 28 August 3:02pm:
"Some quick points:
We are certainly trying to hit the big targets. A priority is stabiltiy amid growth. Today—finally—we should have a dedicated sysadmin signed up, although he will continue to work for his old job for four weeks. That this fellow lives in Australia—Tasmania, in fact—is a nice clue to how hard it's been to find the right person.
Some updates:
1. Better search—general and touchstones—was a big request, and deeply tied into stability. The old search had us keeping two servers just for search. We changed the underlying plumbing, moving from MyISAM to Solr. It's now on a fraction of a server. So, we made progress there.
2. Site speed has also been a concern for us. We've made big strides there, with a sturdy new database and memory caching system. Lots of tag pages have gone from 90 seconds to less than 1.
A lot of this work goes on behind the scenes. For example, I've spent the last two days adding tools to monitor slow pages (rather than database requests). It's given me a guide to priorities.
3. Getting diacriticals right is also a big request. The new search does that, but more generally, Casey is working on a new script to import library data, which will avoid all these problems. It's not for no reason that LibraryThing's competitors don't tie into libraries, let alone parse something like "DANMARC" (the Danish flavor of MARC), or that the Library of Congress's online catalog is almost as fraught with diacritical problems as LT is. We're not dealing with Amazon's XML here.
Best of all, however, hiring Casey and having him work on this stuff means we'll be able to add dozens or more new libraries. Although no single library gets requested over and over, my inbox is FULL of requests to add Hebrew data, or get the Polish one working, etc.
4. Facebook is certainly a big request—while being totally unwanted by another group, but that's life. It's something we have to do to stay competitive, and it takes a LOT of work. We're making progress, but we also recognize a killer Facebook ap. presupposes both collections and a new, generalized data-import function.
5. Collections are, in fact, finally coming. Chris has the underlying code completed, so it's a matter of making a good UI and drawing on the code here and there, as needed. I'm guessing we get it out soon. Soon means this week or next. It'll probably come in slowly.
When you experiment, you go down blind alleys. Collections, for example, went very far down one road—when Chris was working for us some time ago. We had a solution that made collections a special instance of tagging. Going down that road very far convinced us that the underlying idea was flawed and needed to be rethought. Then Christopher, who had written all the code, had to quit LT for a while to finish his BA. It's possible that when someone leaves Microsoft, the dancing paper clip continues to dance. Not on LibraryThing.
Collection is, incidentally, a good example of something that seems easy, but isn't. It basically inserts a new join table into every single script that accesses your books.
6. Beta means fun. You may not like LibraryThing to play, but we're not going to stop. It's good press. Tag mirrors got picked up by some of the top tech blogs, like Weinberger, and the Unsuggester has been a HUGE draw to LibraryThing--from Slashdot to articles in various papers. But we don't do stuff like that because it's good PR. We do it because it's fun. No apologies."
----------------------------------------------------------------------------
Ah, and pleas don't comment his posts here. Go to where the link leads you and comment there.
And If you want to leave a comment about this thread do it in about "Really stalking Tim" rather than here.
So here it starts with the one of Tim's post I stumbled upon today
____________________________________________________________
Tuesday 28 August 3:02pm:
"Some quick points:
We are certainly trying to hit the big targets. A priority is stabiltiy amid growth. Today—finally—we should have a dedicated sysadmin signed up, although he will continue to work for his old job for four weeks. That this fellow lives in Australia—Tasmania, in fact—is a nice clue to how hard it's been to find the right person.
Some updates:
1. Better search—general and touchstones—was a big request, and deeply tied into stability. The old search had us keeping two servers just for search. We changed the underlying plumbing, moving from MyISAM to Solr. It's now on a fraction of a server. So, we made progress there.
2. Site speed has also been a concern for us. We've made big strides there, with a sturdy new database and memory caching system. Lots of tag pages have gone from 90 seconds to less than 1.
A lot of this work goes on behind the scenes. For example, I've spent the last two days adding tools to monitor slow pages (rather than database requests). It's given me a guide to priorities.
3. Getting diacriticals right is also a big request. The new search does that, but more generally, Casey is working on a new script to import library data, which will avoid all these problems. It's not for no reason that LibraryThing's competitors don't tie into libraries, let alone parse something like "DANMARC" (the Danish flavor of MARC), or that the Library of Congress's online catalog is almost as fraught with diacritical problems as LT is. We're not dealing with Amazon's XML here.
Best of all, however, hiring Casey and having him work on this stuff means we'll be able to add dozens or more new libraries. Although no single library gets requested over and over, my inbox is FULL of requests to add Hebrew data, or get the Polish one working, etc.
4. Facebook is certainly a big request—while being totally unwanted by another group, but that's life. It's something we have to do to stay competitive, and it takes a LOT of work. We're making progress, but we also recognize a killer Facebook ap. presupposes both collections and a new, generalized data-import function.
5. Collections are, in fact, finally coming. Chris has the underlying code completed, so it's a matter of making a good UI and drawing on the code here and there, as needed. I'm guessing we get it out soon. Soon means this week or next. It'll probably come in slowly.
When you experiment, you go down blind alleys. Collections, for example, went very far down one road—when Chris was working for us some time ago. We had a solution that made collections a special instance of tagging. Going down that road very far convinced us that the underlying idea was flawed and needed to be rethought. Then Christopher, who had written all the code, had to quit LT for a while to finish his BA. It's possible that when someone leaves Microsoft, the dancing paper clip continues to dance. Not on LibraryThing.
Collection is, incidentally, a good example of something that seems easy, but isn't. It basically inserts a new join table into every single script that accesses your books.
6. Beta means fun. You may not like LibraryThing to play, but we're not going to stop. It's good press. Tag mirrors got picked up by some of the top tech blogs, like Weinberger, and the Unsuggester has been a HUGE draw to LibraryThing--from Slashdot to articles in various papers. But we don't do stuff like that because it's good PR. We do it because it's fun. No apologies."
----------------------------------------------------------------------------
2pivox
Tuesday 28 August 3:59pm:
"Yeah. Oh how we've gone back on forth on booleans and etc. Right now, a book will be able to belong to any number of collections. The trick is the UI. When you want to put something in currently reading you don't want to remove it from "your library." When you put something in "wish list" you do probably do. We're probably going to seek refuge in a check-box UI."
------------------------------------------------------
"Yeah. Oh how we've gone back on forth on booleans and etc. Right now, a book will be able to belong to any number of collections. The trick is the UI. When you want to put something in currently reading you don't want to remove it from "your library." When you put something in "wish list" you do probably do. We're probably going to seek refuge in a check-box UI."
------------------------------------------------------
3rjohara
A simple feature on a person's profile page that would link to "Recent Talk posts by this person" would let anyone stalk anyone, and might be a good idea for a new LT feature.
4Morphidae
>1 pivox: I'm not sure what you are asking. That is the entire purpose of this group. To do exactly what you said.
5myshelves
Morph, maybe you should expand the group description to emphasize that the purpose is to post links, not copies of messages, and to formalize the custom of using a new topic for each link? I'd also suggest adding that discussion of LT features and of whatever Tim & staff say elsewhere should be posted elsewhere, not here. (Just suggestions --- I know who is in charge. *smile*)
I'm also at a loss to understand the reason for this topic, and its title. People have been posting links to messages from staff regarding LT features and policies since the group was started. And while "stalking" was a joke (mine, in fact), arising from a remark by Tim in another group, I don't know what to make of "really stalking." I still don't care what Tim likes on his pizza. :-)
I'm also at a loss to understand the reason for this topic, and its title. People have been posting links to messages from staff regarding LT features and policies since the group was started. And while "stalking" was a joke (mine, in fact), arising from a remark by Tim in another group, I don't know what to make of "really stalking." I still don't care what Tim likes on his pizza. :-)
6_Zoe_
Why is the purpose to post links and not copies of messages, though? I think it would be more convenient to have both a link and a copy of the message.