This can work. Your boss cares a lot about what you deliver, but you have control over how you deliver it. I've used this technique a lot in my career, with good effect.
This is very important. Your concerns just have to be explained and "sold". Any half-decent manager will get a clue when their technology is second-rate: their customers will tell them.
Management will not start out asking for the perfect system, they have deadlines. However, they also don't want a bunch of crap that will cause them to lose customers.
That's where the refactoring is important. It's rare to be able to throw everything out and start over But every update can be an opportunity to improve the system. And over time, that means more reliable technology, and less programming labor costs, which management pays attention to.
I think PostgreSQL's recent success is great. I used MS Access first (in my Windows server days), then moved to PostgreSQL on RH linux. However, this was a few years ago, and the db couldn't keep up with my web app. So i moved to Oracle out of necessity.
I still use PostgreSQL for small projects, and could use it more in the future as it gets better. Oracle rocks, but it's damn expensive, so it's only good for bigger jobs. Plus, open source is cool.
Actually, it can do harm, with wasted human-hours.
This is the crucial point at which we should support an open authentication system alternative-- before there is an installed user base.
Reverse-engineering a closed system is a lot of work, especially if it is a moving target. Wine has taken forever to get usable, MS Word data format is a pain to get right. Samba has to be continually updated, and even the authors say SMB protocol is crap! Think if SMB file sharing died before it became ubiqutous, we wouldn't have to keep chasing microsoft's tail on that one.
Think if TCP/IP weren't so fully supported, we would all be using NetBEUI 2000 for networking.
Re:Ruby is still too new
on
Why not Ruby?
·
· Score: 2
Overall, i have to agree. I do a lot of Perl. The language is good (which is arguable, i know), but the developer support is great.
I was checking out ruby a bit a few months ago, and my main concern was the lack of a standard database library. There were some database interfaces, but no standard way. I think some developers were working on this, but they have a long way to go before they are at the level of Perl's DBI.
Even is this was available, then we'd have to talk about template modules, Apache integration, etc.
I think Ruby has a future, but it's not a no-brainer choice.
yeah, even if this worked, i'm going to get a table fan to take up 4 times the desk space just so i can look at my cpu, and show how l33t i am to my guests.
Maybe he is sensitive about acknowleging Ruby as a real alternative.
It's clear Ruby is playing in the same space as Perl and Python. And Ruby kinda sits in between the two, but seems more Perl-like.
Diclaimer: i'm a Perl programmer (and Java) mostly.
Python has seemed to gain a lot of developer mindshare from those who are turned off by Perl's warts. However, Python brings its own issues, many of which are described in other threads (indents,licensing)
So Ruby is almost like Perl without the warts, as far as oo-inclined programmers are concerned. Ruby's main issue is installed base and developer support.
Developer support (mindshare) is crucial to the success of any software technology. Perl succeeded partly through merits, partly through the historical lack of good alternatives. Python is fighting uphill, and i'm sure guido doesn't want to fight on two fronts.
I think this is known as a strategic partnership. It is in Oracle's interest to run it's software on a large number of platforms, and also to run on free operating systems. That way, corps can cut some costs on paying for Oracle and solaris/windows. More money goes to Oracle, or a larger percentage of the price of deployment. In effect, the database layer becomes the important platform, not the OS.
However, i'm sure they want to keep charging a premium for their proprietary database and app server software, as long as they can.
Plus, it's a bonus for Ellison to stick it to Gates. It makes sense though, as long as Oracle stays significantly better than the open source alternatives.
I've checked out Ruby a bit, since it looks real clean, not as strict as Java, but more easily OO than Perl.
Ruby seems like the classic great technology: better than most alternatives, but is likely to not reach critical mass.
I'm mainly a Perl programmer, that does object-oriented stuff, influenced by my sometimes Java programming. I've found good ways to do clean Perl programming, but the main problem is that everyone has a different approach for doing OO-perl. You can create your own standards, but it's easy to be lax when there is no compliler enforcement.
So i think a good alternative is to create a Perl pragma that enforces Ruby-style coding conventions, in terms of using built-in objects, exceptions, etc. Basic syntax would still be Perl ($%@;\), but there would be an option for a new dialect of Perl that doesn't require a whole new technology.
It could be the next "use strict"! What do you think, Perl guys an gals? We could learn a few lessons without throwing the baby out with the bathwater.
I use ata/66 with a stock 2.4.3 kernel, no patch. This is on RH 6.2, i don't know if RH 7.1 will have it enabled by default, but i'm guessing it will.
I have HPT366 module enabled in kerkel config, but i saw support for CMD64X too. If you are compliing yr own kernel, you just have to enable it, along w/ a few other things.
Try it and forget 2.2, unless you need a stable server or something like that.
The moral of the story is:
Don't wait 'til it expires. Just change the registrar.
Some places like easydns.com (using OpenSRS) will add your remaining time at NSI to your new domain registration. So even if you are cheap, you can move without losing time already paid for.
They refused to remove me as a tech contact from a domain.
Someone i know put me on a domain handled by BulkRegister, because i used to host a couple of their web sites. When this person started being accused of S*P*A*M (may or may not be true), i was getting some heat since i was listed as tech, even though i had nothing to do w/anything at the time.
I contacted BulkRegsiter.com and said remove me. They said:
Thank you for your inquiry. Due to our registration policy, only the member of BulkRegister.com who originally registered the domain name has the access and responsibility to modify the domain record. Please contact the other contacts in the domain record for more information.
On repeated requests, they still refused. I really don't get this stupid policy. What is the point of being a contact if you have absolutely no ability to do anything?
At least NSI will allow a person to "commit suicide", i.e. remove themselves as a contact.
I recently read Hyperspace, by Michio Kaku, and he says many current physicists studying superstring theory postulate 10 dimensions.
I know at first it sounds like they are smoking something, but Einstein's theory of the fourth dimension (time) was also not obvious when he proposed it.
The basic idea is that complicated physical phenomenon can be analyzed more simply using more dimensions. To use the Einstein example, trying to account for variations in distance and time the closer you get to the speed of light would be screwed up calulating in 3 dimensions, but easy adding a 4th one. You just have to accept that space and time can be interchanged, just like the 3 dimensions we see physically.
I think enumerating use cases can definately be helpful. Doing an actual use case diagram, you could skip, though.
Knowing all the actions your system components are expected to support can be useful to get a grasp of your project's scale. I did this before i knew it was called "use cases", so i thought it was cool that there was a name for it (not that i like the name, or anything).
This is the first i head of Curl, and here's my impression, based on the info at the above links:
Free for non-commercial use, pay whatever they say for commercial use
Basically, in today's environment, this will make it hard to get developer support. Open source tools or at least reliably free for use (Java) are the systems that will get adopted (exception: MS.NET)
Custom client simplifies client-server information sharing, using SGML-like language
Even if orgs want this, they are more likely to just use custom java client and XML. I don't see how there will be any substantial web browser support for this, so it will be just another plugin.
I definately understand the complexity of creating web apps, and they need to be simpler to create. But we should create simple frameworks for existing technology, and improve those platforms. I guess they think this will be some kind of quantum leap, but we'll see.
That's the funniest part of the article: Steve Jobs thinks Linus would be more interested in working on OSX vs working on Linux. The assumption is that he would be more interested in hurting Microsoft, rather than actually having full control in creating an OS that has merits in it's own right. I guess it was worth a try, but purely machiavellian.
I understand companies like Apple, Sun, Oracle wanting to compete with microsoft, but i don't like it when CEOs just are slapping at each other and getting in pissing matches.
Plus, can you imagine Linus working for Jobs?
Re:Upgrading from 1.3.x & preserving configuration
on
Apache 2.0 Goes Beta
·
· Score: 2
Hey, i moved from Raven to mod_ssl last year, when the RSA patent expired. It wasn't hard, since raven is based on mod_ssl.
It's a lot easier to keep up to date now, mod_ssl gets updated within a day or two of a new apache release.
If you are building from source, you just have to figure out how you want to build apache, and get those./configure builds right. Once you got it, it's easy to do it for new releases. For a low to medium traffic site, i do a static apache/mod_perl, and dso the other modules, including mod_ssl.
As for your httpd.conf file, it should be the same, except for replacing all the ravenssl stuff with mod_ssl. There will be some pain getting it right, but it shouldn't be too hard.
Slashdot Mirror
This can work. Your boss cares a lot about what you deliver, but you have control over how you deliver it. I've used this technique a lot in my career, with good effect.
This is very important. Your concerns just have to be explained and "sold". Any half-decent manager will get a clue when their technology is second-rate: their customers will tell them.
Management will not start out asking for the perfect system, they have deadlines. However, they also don't want a bunch of crap that will cause them to lose customers.
That's where the refactoring is important. It's rare to be able to throw everything out and start over But every update can be an opportunity to improve the system. And over time, that means more reliable technology, and less programming labor costs, which management pays attention to.
I think PostgreSQL's recent success is great. I used MS Access first (in my Windows server days), then moved to PostgreSQL on RH linux. However, this was a few years ago, and the db couldn't keep up with my web app. So i moved to Oracle out of necessity.
I still use PostgreSQL for small projects, and could use it more in the future as it gets better. Oracle rocks, but it's damn expensive, so it's only good for bigger jobs. Plus, open source is cool.
Actually, it can do harm, with wasted human-hours.
This is the crucial point at which we should support an open authentication system alternative-- before there is an installed user base.
Reverse-engineering a closed system is a lot of work, especially if it is a moving target. Wine has taken forever to get usable, MS Word data format is a pain to get right. Samba has to be continually updated, and even the authors say SMB protocol is crap! Think if SMB file sharing died before it became ubiqutous, we wouldn't have to keep chasing microsoft's tail on that one.
Think if TCP/IP weren't so fully supported, we would all be using NetBEUI 2000 for networking.
Overall, i have to agree. I do a lot of Perl. The language is good (which is arguable, i know), but the developer support is great.
I was checking out ruby a bit a few months ago, and my main concern was the lack of a standard database library. There were some database interfaces, but no standard way. I think some developers were working on this, but they have a long way to go before they are at the level of Perl's DBI.
Even is this was available, then we'd have to talk about template modules, Apache integration, etc.
I think Ruby has a future, but it's not a no-brainer choice.
Uh oh, here we go again.
Just after we did Guido.
We're going to have to start using our extra mod points to hand out Overrateds (but not me, please).
yeah, even if this worked, i'm going to get a table fan to take up 4 times the desk space just so i can look at my cpu, and show how l33t i am to my guests.
Well, i used to do this.
However, i know you can possibly get heating issues w/an open case. Proper airflow usually requires a closed case.
Maybe he is sensitive about acknowleging Ruby as a real alternative.
It's clear Ruby is playing in the same space as Perl and Python. And Ruby kinda sits in between the two, but seems more Perl-like.
Diclaimer: i'm a Perl programmer (and Java) mostly.
Python has seemed to gain a lot of developer mindshare from those who are turned off by Perl's warts. However, Python brings its own issues, many of which are described in other threads (indents,licensing)
So Ruby is almost like Perl without the warts, as far as oo-inclined programmers are concerned. Ruby's main issue is installed base and developer support.
Developer support (mindshare) is crucial to the success of any software technology. Perl succeeded partly through merits, partly through the historical lack of good alternatives. Python is fighting uphill, and i'm sure guido doesn't want to fight on two fronts.
Aren't tabs vs. spaces the reason the makefile format is universally hated?
I think this is known as a strategic partnership. It is in Oracle's interest to run it's software on a large number of platforms, and also to run on free operating systems. That way, corps can cut some costs on paying for Oracle and solaris/windows. More money goes to Oracle, or a larger percentage of the price of deployment. In effect, the database layer becomes the important platform, not the OS.
However, i'm sure they want to keep charging a premium for their proprietary database and app server software, as long as they can.
Plus, it's a bonus for Ellison to stick it to Gates. It makes sense though, as long as Oracle stays significantly better than the open source alternatives.
It's already part of Perl, so why change it?
If you know Perl, is makes sense. If you change it, it's not Perl anymore, then we're back to the issue of everyone using a new language.
I've checked out Ruby a bit, since it looks real clean, not as strict as Java, but more easily OO than Perl.
Ruby seems like the classic great technology: better than most alternatives, but is likely to not reach critical mass.
I'm mainly a Perl programmer, that does object-oriented stuff, influenced by my sometimes Java programming. I've found good ways to do clean Perl programming, but the main problem is that everyone has a different approach for doing OO-perl. You can create your own standards, but it's easy to be lax when there is no compliler enforcement.
So i think a good alternative is to create a Perl pragma that enforces Ruby-style coding conventions, in terms of using built-in objects, exceptions, etc. Basic syntax would still be Perl ($%@;\), but there would be an option for a new dialect of Perl that doesn't require a whole new technology.
It could be the next "use strict"! What do you think, Perl guys an gals? We could learn a few lessons without throwing the baby out with the bathwater.
Mozilla Mail/News works too, i use it w/ IMAP over SSL.
I use ata/66 with a stock 2.4.3 kernel, no patch. This is on RH 6.2, i don't know if RH 7.1 will have it enabled by default, but i'm guessing it will.
I have HPT366 module enabled in kerkel config, but i saw support for CMD64X too. If you are compliing yr own kernel, you just have to enable it, along w/ a few other things.
Try it and forget 2.2, unless you need a stable server or something like that.
This is just an effort for the geeks to buy an Xbox and get the business tax deduction on it!
The moral of the story is:
Don't wait 'til it expires. Just change the registrar.
Some places like easydns.com (using OpenSRS) will add your remaining time at NSI to your new domain registration. So even if you are cheap, you can move without losing time already paid for.
They refused to remove me as a tech contact from a domain.
Someone i know put me on a domain handled by BulkRegister, because i used to host a couple of their web sites. When this person started being accused of S*P*A*M (may or may not be true), i was getting some heat since i was listed as tech, even though i had nothing to do w/anything at the time.
I contacted BulkRegsiter.com and said remove me. They said:
Thank you for your inquiry. Due to our registration policy, only the member of BulkRegister.com who originally registered the domain name has the access and responsibility to modify the domain record. Please contact the other contacts in the domain record for more information.
On repeated requests, they still refused. I really don't get this stupid policy. What is the point of being a contact if you have absolutely no ability to do anything?
At least NSI will allow a person to "commit suicide", i.e. remove themselves as a contact.
Waiting for issue #20!
If it doesn't come out soon, i will start throwing cups at the wall!
I recently read Hyperspace, by Michio Kaku, and he says many current physicists studying superstring theory postulate 10 dimensions.
I know at first it sounds like they are smoking something, but Einstein's theory of the fourth dimension (time) was also not obvious when he proposed it.
The basic idea is that complicated physical phenomenon can be analyzed more simply using more dimensions. To use the Einstein example, trying to account for variations in distance and time the closer you get to the speed of light would be screwed up calulating in 3 dimensions, but easy adding a 4th one. You just have to accept that space and time can be interchanged, just like the 3 dimensions we see physically.
http://www.mkaku.org/hyper_toe.htm
I think enumerating use cases can definately be helpful. Doing an actual use case diagram, you could skip, though.
Knowing all the actions your system components are expected to support can be useful to get a grasp of your project's scale. I did this before i knew it was called "use cases", so i thought it was cool that there was a name for it (not that i like the name, or anything).
Free for non-commercial use, pay whatever they say for commercial use
.NET)
Basically, in today's environment, this will make it hard to get developer support. Open source tools or at least reliably free for use (Java) are the systems that will get adopted (exception: MS
Custom client simplifies client-server information sharing, using SGML-like language
Even if orgs want this, they are more likely to just use custom java client and XML. I don't see how there will be any substantial web browser support for this, so it will be just another plugin.
I definately understand the complexity of creating web apps, and they need to be simpler to create. But we should create simple frameworks for existing technology, and improve those platforms. I guess they think this will be some kind of quantum leap, but we'll see.
That's the funniest part of the article: Steve Jobs thinks Linus would be more interested in working on OSX vs working on Linux. The assumption is that he would be more interested in hurting Microsoft, rather than actually having full control in creating an OS that has merits in it's own right. I guess it was worth a try, but purely machiavellian.
I understand companies like Apple, Sun, Oracle wanting to compete with microsoft, but i don't like it when CEOs just are slapping at each other and getting in pissing matches.
Plus, can you imagine Linus working for Jobs?
Hey, i moved from Raven to mod_ssl last year, when the RSA patent expired. It wasn't hard, since raven is based on mod_ssl.
./configure builds right. Once you got it, it's easy to do it for new releases. For a low to medium traffic site, i do a static apache/mod_perl, and dso the other modules, including mod_ssl.
It's a lot easier to keep up to date now, mod_ssl gets updated within a day or two of a new apache release.
If you are building from source, you just have to figure out how you want to build apache, and get those
As for your httpd.conf file, it should be the same, except for replacing all the ravenssl stuff with mod_ssl. There will be some pain getting it right, but it shouldn't be too hard.