Ideally administrators should aim to automate everything with scripts and then document the use of the scripts.
That way, if you're hit by a bus the whole system is already in a good shape for others to continue. Personally I think it's part of the administrators
responsibility to keep the system in such a good shape. Management should also understand the importance of this and allocate enough time
to keep systems in good shape. Doing things like that as an afterthought takes a really long time and makes knowledge transfers really painful.
If your management has allocated only one day for the knowledge transfer, they're taking a huge risk. There's no way
you can teach everything about a system in a day. What you can do, is tell your current management realistically about the situation and the risks.
After you've gone through the system with the replacement talk to the management about how the knowledge transfer went and make a signed contract
about how much support you are willing to give afterwards and how much it will cost them.
With a contract it's up to the management to decide on how much their system is worth to them and you'll get compensation on doing extra work. Who knows, maybe the management will realise that spending a week on the knowledge transfer upfront might be cheaper than paying you afterwards.
It's nice to see that kind of a reaction in Finland, usually things are way too gray here.
The comments different sides of the debate have been hilarious. The regular church members are blaming the extremists, "You're alienating our members with your extreme views". The extremists are saying "what? we did nothing wrong, that's what is said in the bible, it's not our fault people are leaving". The ultra-extremists are blaming the extremists for not taking a strong enough stance. Public figures and celebrities are taking pot shots at the church and announcing their discontent. Magazines are publishing news about the record resignations.
+1 on distrust and micromanagement, I would say your boss pushing in wrong direction.
If you work on a noisy environment, you could try to prove to your boss that listening to music is actually what shields you from the other sounds & distractions.
a) try to loan a sound level meter from somewhere, preferably a-weighted. There's probably statistics on what level of background noise distracts and chances are your environment produces enough.
b) invite your boss to your cubicle farm to work there for a while. It'll give him the concrete chance to follow on how concentrated you really are on your job and also a possibility to see first hand how distracting your environment is.
Also, based on the possible distrust feelings your boss has, you could try to make your work more visible to him. Once he gets better grasp on your reality, he probably starts feeling more trusting.
Agile process methods like Scrum increase the visibility on how the work is progressing. Scrum also gives more freedom to the team which gets to decide on how it works. In a Scrum process team should also be shielded from external forces, like your boss, so that they can't distract the team. More visibility & freedom -> everyone's happier.
http://en.wikipedia.org/wiki/A-weightinghttp://en.wikipedia.org/wiki/Scrum_(development)
I think the original metric is based on the sign of the WTFs. Positive WTFs are rare and usually the metric goes below zero, sometimes way below;-)
Kidding aside, I'm just reading the book (chapter 2 - Meaningful Names) and from what I've read so far, the book seems really good and useful. In the beginning the author states that some of the guidelines might not work for everyone but at least the author gives a lot of food for thought.
In single player games persistency is generally easy to implement. Most games are based on linear plots, where the gamer can't go backwards. If they can visit the places they've somehow changed, it's also pretty straightforward to save the state of things. No problems in that area.
In multi-user games it isn't so simple anymore. Since most game content (plot, tasks, quests, NPCs etc) is still generated by humans, there's a limit to the content that a single game can contain. If one player completes all the content in the game, what's left for other players? The quests must be somehow reverted back to their initial state. Like in the example, the princess has to be returned back the castle so another player can save her. The easiest way to implement this is by reseting the state after a while. Handling the reset gracefully is the difficult part.
In the example's princess case, graceful reseting might be that the evil king kidnaps the princess again, maybe with the help of a player representing the "other side". Designing quests like this takes more time and resources than the naive "reset after a while"-approach, but maybe we'll start seeing games that behave more naturally.
What you do is you take their comments, and edit them, to make them say exactly the opposite of what they are saying. So, if they say
Rob Sucks!
You can edit it to say
Rob did a great job.
That might work, but it might also be illegal in some countries. Having seen how much time and energy some persons use for pestering and trolling others, I definitely wouldn't want to give a troll legal weapons against me.
A few years ago I had a revelation about email. The revelation was that email is a good tool with a very limited scope. Two persons sending messages is good, three persons is already stretching the limits.
I think that email is definitely not suitable for
task management
planning
storage
version control
communication within a large group
instant messaging
A lot of people try to misuse email for things it's not suitable. Proper tools can make things a lot easier and offer features which email can't offer, ever.
A proper task management software works much better for task related communication (and you get to use the management features at the same time)
Version control systems are a much better place for storing updated documents
Like stated in many replies, GMail conversations >>> email reply chains
Sometimes instant messaging is better than emails
Personal meetings are usually better for planning
Wiki for documentation
...
At the moment the popular applications still seem a little immature because there isn't enough integration between them. Users have to manually move information from application to application which makes life difficult. For example, how many document editors come with an easy to use integrated version control system support?
utilize the power of deterministic destruction, that C# and Java lack, to arrange it so that resources, including but not limited to just memory, are auto-freed. That is incorrect. Java has had WeakReference since 1.2 release which was in 1998.
Weak reference objects, which do not prevent their referents from being made finalizable, finalized, and then reclaimed. The catch is that all code registering disposable objects like events and listeners should use references or objects which use refences, like the WeakHashMap. Sometimes getting that kind automation to work _right_ requires a bit of thought though but abstracting the functionality behind a suitable interface makes things easier to figure out. A generic Eventbroker which uses a Map to hold eventlisteners would enable easy switching between gc'ed (WeakHashMap) and non-gc'ed (any other Map) functionality.
This is a nice example of combining different tools to produce a working solution. Auto-update feature would be required though, because the process seems slightly broken and will loose parts of the data.
He splits the bz2 file into 900kb blocks. The original XML's tags might get broken when a start tag ends up in different block then the end tag. When run, the bash script will ignore all the broken tags.
Fixing that is probably pretty straightforward, but requires a bit more careful XML handling. Anyways, a nice effort and made me want to try the same in a different language:)
I'm not in any hurry to move to Vista or a newer computer. The good time to upgrade might be when SP2 comes out and hardware & software "features" have been ironed out.
OpenFile's OF_SHARE_DENY_NONE flag seems to do allow for such behaviour, but programs rarely seem to use it. I can't remember any situation where I was able to rename an opened file, even when the program has opened it as read only.
+1 for incremential updates
Also, it'd be nice if the user could specify which sub-packages they want. Selecting parts as "I want these three, but not that one" would be very nice for software which has lots of addons.
I've noticed that email works well in some cases. When you need to contact someone about something which is not that important, email is excellent. Also if you want to discuss something important, email is a good media because you can craft your response as long as it takes. The conversation is stored so there's no need to keep separate notes.
Then are those cases when email doesn't work at all, most related to work in some way or the other.
A lot of detail is lost in email. Writing things down is a much slower process and more error prone then describing them in person. There's also the danger of the recipient misunderstanding the message, even to the point where a wrong wording might be understood as an personal insult or challenge.
Email is flooding the inbox. About 20% of the email I get is something that conserns me. Yet I need to constantly filter all the emails in case I miss something important. I'd prefer to read the general announcements from the intranet. No, I'm not interested in mainframe upgrade problems.
Emails get forgotten. "Oh I forgot to send you that email two weeks ago"
Discussions about complex and large issues usually get derailed if there are too many people chipping in. Details are lost in the flood of the messages as people fork their own discussions.... and so on.
Finding out good ways to reduce the amount of emails isn't easy. It seems that the best principle is to use the right tool for the job.
Bug and issue tracking, tasks and project management work much better when a proper project management software is used. Concentrating on the right tasks is much easier, if you just need to check from the application what you're supposed to be doing. Everything gets assigned to a someone so it's always obvious who should be working on what. This is an interesting piece about how Linden Labs uses a task management software to control their every day tasks -> http://www.atlassian.com/software/jira/casestudies /linden.jsp "We use JIRA for all task-in-progress tracking in the company - everything from ordering food in the kitchen to releasing a new feature to fixing a bug."
Face to face discussions, meetings (in moderate amount) and pair programming increase the flow of information. Forums work nicely for discussions in a larger group. Wikis are good for sharing information. Somehow I just wish they'd all come as a nicely integrated package which allows the information flow as naturally as possible. (and no, Lotus Notes is not integrated, nice or natural)
Keeping user files on a separate partition of their own helps if you ever need to install a new os. That way you don't have move user files to a safe place. You can even map user directories like "My Documents" to the user partition's directories so you'll get nice defaults for file save locations.
I've used JIRA at two different employers and it's a very nice product. The interface is powerful and at the same time simple enough (looking at you bugzilla). It works with different databases, but it also has a standalone install which a breeze to install. Comes with nice integration to cvs ( & subversion too I think) and (paying?) customers get the source code too so you can integrate however you like.
The only downside to Jira is that it's price tag (for business users) has risen steadily, but at least they've given free licenses to open source projects like Apache Software Foundation, Codehaus and JBoss.
I've also used Mercury's TestDirector, but it seems like a glorified excel-sheet when compared to JIRA. TD is more suitable for reporting bugs, and it doesn't support the software development process like JIRA does. Jira has projects, components, issue links, releases, change notes, workflows, security levels, reports and so on.
We also have an inhouse built issue tracking system. It works to some extent, but its GUI doesn't really scale to handling large number of tickets. And since it's not developed actively it will probably stay as it is for some time.
ArrayList, LinkedList and Vector implement the List interface. Collections.synchronizedList(List list) returns a synchronized (thread-safe) list backed by the specified list, so there's also a generic way of "making" synchronized lists. The same method (Collections.synchronized*) is also implemented for maps & sets.
Imo the array primitive type could have been an object instead of the current primitive type implementation but I don't see what's wrong with using the ArrayList, since it's pretty close to an Array object anyway. You just don't get to use array's syntax which isn't that big of a thing with modern IDEs.
With Java's lists, sets and maps you aren't using just "language-provided array objects", the class names reflect the way the objects function. I haven't used BLAS (or FORTRAN) but according to google it's "A high quality "building block" routines for performing basic vector and matrix operations." where at least the vector part sounds awfully close to ArrayList and its kind;)
Do you mean arraylist, vector, arrays and such? Different implementations are supposed to be used in different scenarios.
arrays = non-dynamic, very fast insertion & seek
ArrayList = dynamic, moderate memory usage, quick seek, possibly slow remove & insertion
LinkedList = dynamic, quick remove & insertion, slow seek & uses more memory than arraylist
There's also different algorithms & implementations for different concurrency needs & so on.
Having abstract interfaces to different implementations & algorithms is definitely a nice thing. You can implement your code by using the any implementations. You need to worry about the different implementations only if your code is not efficient enough and based on profiling and analysis you should be able to determine which implementation(s) should be used.
It's pretty hard to say if it's due the greenhouse effect, but Finland hasn't gotten permanent snow yet (at least in the southern parts). I remember a time when I had to walk through snow banks on my birthday (7th of oct) but this year it's almost warm enough to cycle to work in shorts (+7C).
The weather is breaking temperature records and even in the northern parts it's not better. For example tourism is getting hit badly. Yesterday I saw some news footage where tourists were walking in something which was much closer to water then snow and they were supposed to be on a skiing holiday. And even worse, if it doesn't snow the Finnish Santa isn't going to get off the ground in two weeks.
I use XStream which is very nice xml serialization utility library. XStream has very nice aliasing abilities, so you can trim the xml to very close to absolute minimum ( 2 minute tutorial).
Currently I'm planning on externalizing website validation objects as xml so they can be updated more easily and dynamically. It's basically the same thing you describe and I can easily see how it would help with GUI programming. To get around the "attaching actions to the components" problem you can use the MVC design pattern. Set the gui elements to return an event id and use a central controller class to send the event flow to the proper commands.
The XML-serialization has a few problems though:
Versioning, what if the classes change and what if the data model changes really radically.
XML data doesn't get refactored when you refactor your classes.
XML data might not stay in synch with your codebase.
The XML becomes easily convoluted and at worst you'll be coding in XML without the help of a friendly IDE.
Anyways, XSLT can be used to convert older XML formats to new format and XStream's aliasing functions helps, but you still have to be careful when making changes.
Slashdot effect * pipelining with 20 connections = 20 times faster slashdotting.
I think you're right about the 2 connections being a little 1990s, but over eight seems a little excessive. And hitting HTTP servers with 20 simultaneous connections is a bit nasty. Instead of serving nice clients with two connections each, the server might allocate ten times more resources to serve the greedy one. Also, "keep-alive" keeps the connections alive for a while (FF default: 300 (seconds?)) so the server might keep holding on to the resources much longer than that one page load.
Hitting an idle server won't probably have any effect, but slashdotting with 20 connections per client is much faster than with 2 per client.
Here in Finland we just had a judgement in the case about a 10000 user torrent site called Finreactor. 22 defendants have to pay 420 000 in damages and 142 000 in court fees.
Considering normal Finnish judgements, the payments are unusually high, even when compared to crimes like rape or murder. The defendants will probably appeal to a higher court so the case isn't over yet.
To make matters even worse, there's a new copyright law in effect. This case was judged by the older, more lenient copyright law. The new law is even more strict than the EU guideline it's based on and other countries have implemented much nicer versions. The Finnish version goes ridiculous lengths to assure that pirates are punished. There's even a clause which allows the copyright owner to demand that the guilty party publishes an announcement about the judgement.
There was a huge (in Finnish standards anyway) public outcry about the law and its loop holes but to no avail. The record industry lobbyists were so successful in getting the trust of the law makers that one of them actually wrote in her public blog about how "it was nice that the men from Teosto were thanking me so much about the law after the nasty public outcry". She was at a party organised by the record industry no less.
Slashdot Mirror
Ideally administrators should aim to automate everything with scripts and then document the use of the scripts. That way, if you're hit by a bus the whole system is already in a good shape for others to continue. Personally I think it's part of the administrators responsibility to keep the system in such a good shape. Management should also understand the importance of this and allocate enough time to keep systems in good shape. Doing things like that as an afterthought takes a really long time and makes knowledge transfers really painful.
If your management has allocated only one day for the knowledge transfer, they're taking a huge risk. There's no way you can teach everything about a system in a day. What you can do, is tell your current management realistically about the situation and the risks. After you've gone through the system with the replacement talk to the management about how the knowledge transfer went and make a signed contract about how much support you are willing to give afterwards and how much it will cost them.
With a contract it's up to the management to decide on how much their system is worth to them and you'll get compensation on doing extra work. Who knows, maybe the management will realise that spending a week on the knowledge transfer upfront might be cheaper than paying you afterwards.
Here are graphs showing the increase:
http://eroakirkosta.fi/static/ek-tilastot/vuodet.png
http://eroakirkosta.fi/static/ek-tilastot/
It's nice to see that kind of a reaction in Finland, usually things are way too gray here.
The comments different sides of the debate have been hilarious. The regular church members are blaming the extremists, "You're alienating our members with your extreme views". The extremists are saying "what? we did nothing wrong, that's what is said in the bible, it's not our fault people are leaving". The ultra-extremists are blaming the extremists for not taking a strong enough stance. Public figures and celebrities are taking pot shots at the church and announcing their discontent. Magazines are publishing news about the record resignations.
I'm thinking about getting HP ZR24W, it's a quality screen with 1920x1200. At least it won't be a setback from 1600x1200.
http://www.tftcentral.co.uk/reviews/hp_zr24w.htm
+1 on distrust and micromanagement, I would say your boss pushing in wrong direction. If you work on a noisy environment, you could try to prove to your boss that listening to music is actually what shields you from the other sounds & distractions. a) try to loan a sound level meter from somewhere, preferably a-weighted. There's probably statistics on what level of background noise distracts and chances are your environment produces enough. b) invite your boss to your cubicle farm to work there for a while. It'll give him the concrete chance to follow on how concentrated you really are on your job and also a possibility to see first hand how distracting your environment is. Also, based on the possible distrust feelings your boss has, you could try to make your work more visible to him. Once he gets better grasp on your reality, he probably starts feeling more trusting. Agile process methods like Scrum increase the visibility on how the work is progressing. Scrum also gives more freedom to the team which gets to decide on how it works. In a Scrum process team should also be shielded from external forces, like your boss, so that they can't distract the team. More visibility & freedom -> everyone's happier. http://en.wikipedia.org/wiki/A-weighting http://en.wikipedia.org/wiki/Scrum_(development)
I think the original metric is based on the sign of the WTFs. Positive WTFs are rare and usually the metric goes below zero, sometimes way below ;-)
Kidding aside, I'm just reading the book (chapter 2 - Meaningful Names) and from what I've read so far, the book seems really good and useful. In the beginning the author states that some of the guidelines might not work for everyone but at least the author gives a lot of food for thought.
In multi-user games it isn't so simple anymore. Since most game content (plot, tasks, quests, NPCs etc) is still generated by humans, there's a limit to the content that a single game can contain. If one player completes all the content in the game, what's left for other players? The quests must be somehow reverted back to their initial state. Like in the example, the princess has to be returned back the castle so another player can save her. The easiest way to implement this is by reseting the state after a while. Handling the reset gracefully is the difficult part.
In the example's princess case, graceful reseting might be that the evil king kidnaps the princess again, maybe with the help of a player representing the "other side". Designing quests like this takes more time and resources than the naive "reset after a while"-approach, but maybe we'll start seeing games that behave more naturally.
What you do is you take their comments, and edit them, to make them say exactly the opposite of what they are saying. So, if they say
Rob Sucks!
You can edit it to say
Rob did a great job.
That might work, but it might also be illegal in some countries. Having seen how much time and energy some persons use for pestering and trolling others, I definitely wouldn't want to give a troll legal weapons against me.
I think that email is definitely not suitable for
- task management
- planning
- storage
- version control
- communication within a large group
- instant messaging
A lot of people try to misuse email for things it's not suitable. Proper tools can make things a lot easier and offer features which email can't offer, ever.- A proper task management software works much better for task related communication (and you get to use the management features at the same time)
- Version control systems are a much better place for storing updated documents
- Like stated in many replies, GMail conversations >>> email reply chains
- Sometimes instant messaging is better than emails
- Personal meetings are usually better for planning
- Wiki for documentation
...
At the moment the popular applications still seem a little immature because there isn't enough integration between them. Users have to manually move information from application to application which makes life difficult. For example, how many document editors come with an easy to use integrated version control system support?He splits the bz2 file into 900kb blocks. The original XML's tags might get broken when a start tag ends up in different block then the end tag. When run, the bash script will ignore all the broken tags.
Fixing that is probably pretty straightforward, but requires a bit more careful XML handling. Anyways, a nice effort and made me want to try the same in a different language :)
I'm not in any hurry to move to Vista or a newer computer. The good time to upgrade might be when SP2 comes out and hardware & software "features" have been ironed out.
OpenFile's OF_SHARE_DENY_NONE flag seems to do allow for such behaviour, but programs rarely seem to use it. I can't remember any situation where I was able to rename an opened file, even when the program has opened it as read only.
How do you rename something which is in use? Windows keeps it usually locked.
+1 for incremential updates Also, it'd be nice if the user could specify which sub-packages they want. Selecting parts as "I want these three, but not that one" would be very nice for software which has lots of addons.
I've noticed that email works well in some cases. When you need to contact someone about something which is not that important, email is excellent. Also if you want to discuss something important, email is a good media because you can craft your response as long as it takes. The conversation is stored so there's no need to keep separate notes.
... and so on.
s /linden.jsp "We use JIRA for all task-in-progress tracking in the company - everything from ordering food in the kitchen to releasing a new feature to fixing a bug."
Then are those cases when email doesn't work at all, most related to work in some way or the other.
A lot of detail is lost in email. Writing things down is a much slower process and more error prone then describing them in person. There's also the danger of the recipient misunderstanding the message, even to the point where a wrong wording might be understood as an personal insult or challenge.
Email is flooding the inbox. About 20% of the email I get is something that conserns me. Yet I need to constantly filter all the emails in case I miss something important. I'd prefer to read the general announcements from the intranet. No, I'm not interested in mainframe upgrade problems.
Emails get forgotten. "Oh I forgot to send you that email two weeks ago"
Discussions about complex and large issues usually get derailed if there are too many people chipping in. Details are lost in the flood of the messages as people fork their own discussions.
Finding out good ways to reduce the amount of emails isn't easy. It seems that the best principle is to use the right tool for the job.
Bug and issue tracking, tasks and project management work much better when a proper project management software is used. Concentrating on the right tasks is much easier, if you just need to check from the application what you're supposed to be doing. Everything gets assigned to a someone so it's always obvious who should be working on what. This is an interesting piece about how Linden Labs uses a task management software to control their every day tasks -> http://www.atlassian.com/software/jira/casestudie
Face to face discussions, meetings (in moderate amount) and pair programming increase the flow of information. Forums work nicely for discussions in a larger group. Wikis are good for sharing information. Somehow I just wish they'd all come as a nicely integrated package which allows the information flow as naturally as possible. (and no, Lotus Notes is not integrated, nice or natural)
Keeping user files on a separate partition of their own helps if you ever need to install a new os. That way you don't have move user files to a safe place. You can even map user directories like "My Documents" to the user partition's directories so you'll get nice defaults for file save locations.
The only downside to Jira is that it's price tag (for business users) has risen steadily, but at least they've given free licenses to open source projects like Apache Software Foundation, Codehaus and JBoss.
I've also used Mercury's TestDirector, but it seems like a glorified excel-sheet when compared to JIRA. TD is more suitable for reporting bugs, and it doesn't support the software development process like JIRA does. Jira has projects, components, issue links, releases, change notes, workflows, security levels, reports and so on.
We also have an inhouse built issue tracking system. It works to some extent, but its GUI doesn't really scale to handling large number of tickets. And since it's not developed actively it will probably stay as it is for some time.
Imo the array primitive type could have been an object instead of the current primitive type implementation but I don't see what's wrong with using the ArrayList, since it's pretty close to an Array object anyway. You just don't get to use array's syntax which isn't that big of a thing with modern IDEs.
With Java's lists, sets and maps you aren't using just "language-provided array objects", the class names reflect the way the objects function. I haven't used BLAS (or FORTRAN) but according to google it's "A high quality "building block" routines for performing basic vector and matrix operations." where at least the vector part sounds awfully close to ArrayList and its kind ;)
Do you mean arraylist, vector, arrays and such? Different implementations are supposed to be used in different scenarios.
arrays = non-dynamic, very fast insertion & seekArrayList = dynamic, moderate memory usage, quick seek, possibly slow remove & insertion
LinkedList = dynamic, quick remove & insertion, slow seek & uses more memory than arraylist
There's also different algorithms & implementations for different concurrency needs & so on.
Having abstract interfaces to different implementations & algorithms is definitely a nice thing. You can implement your code by using the any implementations. You need to worry about the different implementations only if your code is not efficient enough and based on profiling and analysis you should be able to determine which implementation(s) should be used.
The weather is breaking temperature records and even in the northern parts it's not better. For example tourism is getting hit badly. Yesterday I saw some news footage where tourists were walking in something which was much closer to water then snow and they were supposed to be on a skiing holiday. And even worse, if it doesn't snow the Finnish Santa isn't going to get off the ground in two weeks.
I use XStream which is very nice xml serialization utility library. XStream has very nice aliasing abilities, so you can trim the xml to very close to absolute minimum ( 2 minute tutorial).
Currently I'm planning on externalizing website validation objects as xml so they can be updated more easily and dynamically. It's basically the same thing you describe and I can easily see how it would help with GUI programming. To get around the "attaching actions to the components" problem you can use the MVC design pattern. Set the gui elements to return an event id and use a central controller class to send the event flow to the proper commands.
The XML-serialization has a few problems though:
Anyways, XSLT can be used to convert older XML formats to new format and XStream's aliasing functions helps, but you still have to be careful when making changes.
I think you're right about the 2 connections being a little 1990s, but over eight seems a little excessive. And hitting HTTP servers with 20 simultaneous connections is a bit nasty. Instead of serving nice clients with two connections each, the server might allocate ten times more resources to serve the greedy one. Also, "keep-alive" keeps the connections alive for a while (FF default: 300 (seconds?)) so the server might keep holding on to the resources much longer than that one page load.
Hitting an idle server won't probably have any effect, but slashdotting with 20 connections per client is much faster than with 2 per client.
Considering normal Finnish judgements, the payments are unusually high, even when compared to crimes like rape or murder. The defendants will probably appeal to a higher court so the case isn't over yet.
To make matters even worse, there's a new copyright law in effect. This case was judged by the older, more lenient copyright law. The new law is even more strict than the EU guideline it's based on and other countries have implemented much nicer versions. The Finnish version goes ridiculous lengths to assure that pirates are punished. There's even a clause which allows the copyright owner to demand that the guilty party publishes an announcement about the judgement.
There was a huge (in Finnish standards anyway) public outcry about the law and its loop holes but to no avail. The record industry lobbyists were so successful in getting the trust of the law makers that one of them actually wrote in her public blog about how "it was nice that the men from Teosto were thanking me so much about the law after the nasty public outcry". She was at a party organised by the record industry no less.
One can wish.