We have used Perforce for a group of up to 75 developers, 30000+ files, across 3 geographic sites. Currently we do about a 1000 changes a year on 6 inter-related products. I can attest that it is very robust and is the best product we tried across a WAN. We have been very happy with it, especially since it was fairly affordable (compared to things like ClearCase and Continuus).
Perforce concentrates on doing one thing well - tracking of software changes. It is not intended to be a work tracking system, which some of the more expensive products provide. It does not provide a GUI (although, why would you ever want to leave emacs anyway?). We use it integrated with our own web-based work management system and did not have to adapt our practices to it. The merge capability is excellent.
It's weakness are tolerable and somewhat unobvious. We felt it important to be able to compare changes in aggregate releases for debugging and reporting purposes. We found no products that do this well. The bookkeeping for this is a little tedious in Perforce and somewhat prone to misreporting old changes from deleted files. The form-based interface can also be a little confusing for things like branching - I see no reason why this could not be simplified.
In looking at configuration management, I think you have to concentrate on some of the basic principals:
like always knowing what you have
like being able to monitor changes in the software
like the tool being an aide to getting work done, not a barrier
like the tool supporting your change process, not imposing it's own
It is very difficult to evaluate some these products (mainly due to pushy sales people). Perforce was easy to evaluate and worked well for us. Putting in Perforce (migrating from CVS) was easy. Don't be fooled by a fancy GUI.
Finally some sense on Java
on
Bitter Java
·
· Score: 1
Java has been the source of incredibly hype among in the industry. It is finally nice to hear some sensible discussion about it.
My department has quietly deprecated Java in favour of simpler building blocks (C for big real-time applications, Perl and VB for small apps). Our projects are always on-time and on-budget as a result. In comparison, hard-core OOP languages like C++ and Java have resulted in disasters ($$$$$).
I'm not against them in principal, but the problem with Java (and to a lesser extent C++) is that OOP requires a fair bit of expertise to master. The issues are subtle and deep. These projects I would not something I would give to anyone under 5-10 years of programming experience which unfortunately seems to be the profile of your average Big-5 IT consultant.
Java is still platform-sensitive and vulnerable to run-time errors. Its sole strength is database connectivity. You can implement some OOP concepts in C but structuring your program well. This usually gets you the results you want. OO Perl is very nice, but not for novices.
A final comment - when I see a well made web site it seems like the scripts often end in.pl or.php. I'm starting to see good ones in.asp. Most often, the really pathetic ones end in.jsp. The results speak for themselves.
I can confirm that prices are significantly better up here on many things from cars to Internet access. This seems to be because the market will bear a much higher price in the US for consumer items. That is because Canada has roughly the same affluence as the poorest state in America.
Significantly, real estate is a fair bit cheaper in major cities here than in the U.S. Office space is running about US$10/square foot (I think) and houses start about US$175,000 in Toronto. Taxes are somewhat higher, but you do get some mediocre health care thrown in for free. Yes, the heating bill is nasty in the winter.
Anyway, if you are a business, these things might be a big factor.
The downside (or upside depending on who you are) - from every salary survey I have seen, salaries here are much lower. Hard to get a handle on this, but my impression is that technical wages in U.S. are the same as Canada but in US$ (thus 40% higher?). Part of the problem is the "branch plant" factor in the Canadian economy.
Like, at least our beer is way better, eh?
BTW No, and we don't pronounce it "a-boot" Who started that myth?
The 20k$ price is about right depending on how many documents it indexes. Altavista charges about
34k$ CDN for 50,000 documents (a limit we have blown on just one corporate web site). I have heard that other engines are similarly priced.
This is not including fairly high maintenance fees...
When you consider that any corporate site could be a window on a huge corporate database of information (literally, a each web page could be a database record), you could blow through hundreds of thousands of documents easily.
For the uber-geeks out there, search engines work nice with configuration management systems (like Perforce) for searching source code for large projects.
By the way, my experience with the Altavista product is it is very buggy and unreliable (they re-wrote it in Java, any surprise?), so Google's entry into the field is welcome.
You can't lose customers when you have a monopoly in most areas. Bell Sympatico and Look Communications simply aren't available in
large parts of even major centers like Toronto.
I've got no choice but to pay, although I may write my MP...
In 20 years of using computers, I've never seen so much poor software than in the last 12 months. I'm curious about the video flickering/blacking problem the reviewer mentioned - I observed the same problem in Black & White rendering the game unusable. I should mention this was on PIII/500, not some ancient machine.
Similar examples of low quality software are Nero (failed to burn a single good CD), and Visio 2000 (new computers seem to seg fault beginning with installation). Quicken QuickTax 2000 wouldn't update unless you installed the full-version (extensive debugging required before I could file my taxes online). I've seen at least four examples of software that ignore installation options (like not installing desktop icons, not checking for updates on the internet, etc.)
Thank goodness I generally don't pay for this crap.
It depends, I think on what sort of Webcam you have. I tried a Creative Labs Webcam and pointed
it out my sons window. It was very obvious and
I'm sure any serious theif would just avoid the
area and steal the CPU.
At best all I was able to get was the back of the
mailman walking away because the software triggered to slowly. Increasing the sensitivity
would cause it to trigger on passing cloud shadows or shadows from cars on the street.
It was very difficult to get a good mount point and a decent field of view.
My speed on Rogers Cable at the east end of Toronto is generally 10kB/s to 100kB/s. The real problem is lack of reliability. Today my service is down because the DHCP server is mis-configured. Last month a central router appeared to be overloading. Of course, the cable company will rarely admit to these problems.
Sadly, DSL is not even available in my neighborhood, so the cable monopoly looks great compared to the local phone monopoly. Did I mention the wireless cable/Internet alternative is going bankrupt and overbilled an unknown number of customer's credit cards? Pretty pathetic service options for a city of 2.5 million people.
It seems like most people have missed the point of this article. It rightly asserts that the schooling the computer/software engineer/scientist has a big influence on the suckiness of software. As a software engineer who has been involved in recruiting at universities lately it is pretty tough to find good people.
One problem with Computer Science is students get little education in physics or technology. The more often receive all their computer knowledge from university and have little breadth of experience. Engineers on the other hand generally get only 2 or 3 computer courses, but have broader experience and interesy both inside and outside of school which more than compensates. I would advocate the transfer of CS departments from Math faculties to engineering as a result (this seems to be a growing trend). Note that certification programs like MSCE, etc. carry zero weight with us and are almost a negative.
The teaching of Java in school is particularly troubling. Students who mainly have background in Java often really don't have a depth of understanding on how the computer works. This results in bloated kludges in whatever language they work in.
Example - anybody can put together an HTML web page. But, it requires a real depth of knowledge and breadth of experience to architect a commercial web application which doesn't suck. This 'big picture' view is what I try to hire for.
The track record of the cable and telephone company in Canada on this is terrible. There are only two major cable companies (Shaw and Rogers) and both have terrible records for customer service. Local telephone service is an monopoly in most areas and Bell Canada has been extremely slow to rollout high-speed service.
You don't have to live in a rural area, even in Toronto, high speed service through the phone company is not available in many areas of the city.
Fortunately, wireless microwave Internet (recently introduced) is going to break through the barrier and delivers both video and Internet at higher speeds and cheaper prices than alternatives. Mine gets installed next week - can't wait!
After reading Dune:House Attrides I was fairly impressed as it provided the background material only hinted at in the original. I then proceded to re-read the entire series (for approximately the 21st time, no kidding) and noticed glaring problems in plot continuity throughout the Frank Herbert volumes. Don't get me wrong, they are great science fiction, but they aren't perfect (Tolkien on the other hand, comes very close to perfect).
Frank Herbert's books focus on a few key ideas about cellular memory, prescience and other special abilities. The broad sweep of ideas is what's important, not plotline. The prequels are more pure fun science fiction, and there is nothing wrong with that. Dune: House Harkonnenen is already on my nightstand awaiting the completion of a great Neal Stephenson novel. I'm looking forward to it.
Removable hard drives are standard on high-end equipment. Compaq has two different solutions - the Storageworks stuff that came from Digital is great, we've used it for years and you don't even have to power down the computer for it to recognize the drive. Of course, this stuff is all SCSI, which PC vendors still haven't standardized on.
This is not a good replacement for a floppy though, a hard drive is still $200 and the size is too large. The replacement should have no moving parts.
If you look at the cost/capacity curve of cheap portable media over the last 20 years, the floppy disk is surely an anomoly. For example, one reference point on this curve is the Commodore 4040 hard drive - about $1800 in 1980, two 5 1/4 inches, about 180kB (I think). Very reliable - I still have disks that work.
The 3 1/2 inch floppy has been standard equipment for about 13 years, and really has not advanced any. Last time I bought one it cost me $24.
Obviously some group of vendors needs to get together to research and push a new standard. People are using hard drives as portable media, which is a sad state of affairs. Until the right product is found in your local PC builder's shop, it will not take off.
Soild state media may be do-able because chip prices are so low, but the product would still probably be too expensive to be disposable. It would also have to be extremely strong to survive the abuse I would dish out. Perhaps a 1/2x1x2" thick rectangle with a belt clip, static-protected USB connection at one end, and a LCD readout on other to indicate capacity remaining. Cost should be no more than $40 for 100MB. It must be able to plug in with no drive necessary.
Slashdot Mirror
Perforce concentrates on doing one thing well - tracking of software changes. It is not intended to be a work tracking system, which some of the more expensive products provide. It does not provide a GUI (although, why would you ever want to leave emacs anyway?). We use it integrated with our own web-based work management system and did not have to adapt our practices to it. The merge capability is excellent.
It's weakness are tolerable and somewhat unobvious. We felt it important to be able to compare changes in aggregate releases for debugging and reporting purposes. We found no products that do this well. The bookkeeping for this is a little tedious in Perforce and somewhat prone to misreporting old changes from deleted files. The form-based interface can also be a little confusing for things like branching - I see no reason why this could not be simplified.
In looking at configuration management, I think you have to concentrate on some of the basic principals:
like always knowing what you have
like being able to monitor changes in the software
like the tool being an aide to getting work done, not a barrier
like the tool supporting your change process, not imposing it's own
It is very difficult to evaluate some these products (mainly due to pushy sales people). Perforce was easy to evaluate and worked well for us. Putting in Perforce (migrating from CVS) was easy. Don't be fooled by a fancy GUI.
My department has quietly deprecated Java in favour of simpler building blocks (C for big real-time applications, Perl and VB for small apps). Our projects are always on-time and on-budget as a result. In comparison, hard-core OOP languages like C++ and Java have resulted in disasters ($$$$$).
I'm not against them in principal, but the problem with Java (and to a lesser extent C++) is that OOP requires a fair bit of expertise to master. The issues are subtle and deep. These projects I would not something I would give to anyone under 5-10 years of programming experience which unfortunately seems to be the profile of your average Big-5 IT consultant.
Java is still platform-sensitive and vulnerable to run-time errors. Its sole strength is database connectivity. You can implement some OOP concepts in C but structuring your program well. This usually gets you the results you want. OO Perl is very nice, but not for novices.
A final comment - when I see a well made web site it seems like the scripts often end in .pl or .php. I'm starting to see good ones in .asp. Most often, the really pathetic ones end in .jsp. The results speak for themselves.
Significantly, real estate is a fair bit cheaper in major cities here than in the U.S. Office space is running about US$10/square foot (I think) and houses start about US$175,000 in Toronto. Taxes are somewhat higher, but you do get some mediocre health care thrown in for free. Yes, the heating bill is nasty in the winter.
Anyway, if you are a business, these things might be a big factor.
The downside (or upside depending on who you are) - from every salary survey I have seen, salaries here are much lower. Hard to get a handle on this, but my impression is that technical wages in U.S. are the same as Canada but in US$ (thus 40% higher?). Part of the problem is the "branch plant" factor in the Canadian economy.
Like, at least our beer is way better, eh?
BTW No, and we don't pronounce it "a-boot" Who started that myth?
This is not including fairly high maintenance fees...
When you consider that any corporate site could be a window on a huge corporate database of information (literally, a each web page could be a database record), you could blow through hundreds of thousands of documents easily.
For the uber-geeks out there, search engines work nice with configuration management systems (like Perforce) for searching source code for large projects.
By the way, my experience with the Altavista product is it is very buggy and unreliable (they re-wrote it in Java, any surprise?), so Google's entry into the field is welcome.
I've got no choice but to pay, although I may write my MP...
Now, what happens when Peer-to-Peer and the corporate world really get put together?
Its more likely due to Celine Dion and Sarah McLaughlin being on hiatus than Napster! :-)
Similar examples of low quality software are Nero (failed to burn a single good CD), and Visio 2000 (new computers seem to seg fault beginning with installation). Quicken QuickTax 2000 wouldn't update unless you installed the full-version (extensive debugging required before I could file my taxes online). I've seen at least four examples of software that ignore installation options (like not installing desktop icons, not checking for updates on the internet, etc.)
Thank goodness I generally don't pay for this crap.
It depends, I think on what sort of Webcam you have. I tried a Creative Labs Webcam and pointed it out my sons window. It was very obvious and I'm sure any serious theif would just avoid the area and steal the CPU. At best all I was able to get was the back of the mailman walking away because the software triggered to slowly. Increasing the sensitivity would cause it to trigger on passing cloud shadows or shadows from cars on the street. It was very difficult to get a good mount point and a decent field of view.
Sadly, DSL is not even available in my neighborhood, so the cable monopoly looks great compared to the local phone monopoly. Did I mention the wireless cable/Internet alternative is going bankrupt and overbilled an unknown number of customer's credit cards? Pretty pathetic service options for a city of 2.5 million people.
One problem with Computer Science is students get little education in physics or technology. The more often receive all their computer knowledge from university and have little breadth of experience. Engineers on the other hand generally get only 2 or 3 computer courses, but have broader experience and interesy both inside and outside of school which more than compensates. I would advocate the transfer of CS departments from Math faculties to engineering as a result (this seems to be a growing trend). Note that certification programs like MSCE, etc. carry zero weight with us and are almost a negative.
The teaching of Java in school is particularly troubling. Students who mainly have background in Java often really don't have a depth of understanding on how the computer works. This results in bloated kludges in whatever language they work in.
Example - anybody can put together an HTML web page. But, it requires a real depth of knowledge and breadth of experience to architect a commercial web application which doesn't suck. This 'big picture' view is what I try to hire for.
You don't have to live in a rural area, even in Toronto, high speed service through the phone company is not available in many areas of the city.
Fortunately, wireless microwave Internet (recently introduced) is going to break through the barrier and delivers both video and Internet at higher speeds and cheaper prices than alternatives. Mine gets installed next week - can't wait!
Frank Herbert's books focus on a few key ideas about cellular memory, prescience and other special abilities. The broad sweep of ideas is what's important, not plotline. The prequels are more pure fun science fiction, and there is nothing wrong with that. Dune: House Harkonnenen is already on my nightstand awaiting the completion of a great Neal Stephenson novel. I'm looking forward to it.
Removable hard drives are standard on high-end equipment. Compaq has two different solutions - the Storageworks stuff that came from Digital is great, we've used it for years and you don't even have to power down the computer for it to recognize the drive. Of course, this stuff is all SCSI, which PC vendors still haven't standardized on. This is not a good replacement for a floppy though, a hard drive is still $200 and the size is too large. The replacement should have no moving parts.
If you look at the cost/capacity curve of cheap portable media over the last 20 years, the floppy disk is surely an anomoly. For example, one reference point on this curve is the Commodore 4040 hard drive - about $1800 in 1980, two 5 1/4 inches, about 180kB (I think). Very reliable - I still have disks that work. The 3 1/2 inch floppy has been standard equipment for about 13 years, and really has not advanced any. Last time I bought one it cost me $24. Obviously some group of vendors needs to get together to research and push a new standard. People are using hard drives as portable media, which is a sad state of affairs. Until the right product is found in your local PC builder's shop, it will not take off. Soild state media may be do-able because chip prices are so low, but the product would still probably be too expensive to be disposable. It would also have to be extremely strong to survive the abuse I would dish out. Perhaps a 1/2x1x2" thick rectangle with a belt clip, static-protected USB connection at one end, and a LCD readout on other to indicate capacity remaining. Cost should be no more than $40 for 100MB. It must be able to plug in with no drive necessary.