While this approach would work in certain circumstances, there's still an assumption there that the original system was built to accomodate that sort of scenario. If it was, great. If it wasn't, then you're stuck.
1) Triggers are hidden application logic that are very hard to debug, and are easily overlooked or forgotten by developers. Business logic (other than defensive logic like unique indexes, primary keys, foreign keys, not-null columns) does not belong in the database. They belong in the middle tier. They also make it much more difficult to move to another database.
I used to think as you write here, but have modified my thinking some in the past few years. I agree with your view on the 'hidden application logic', it might not always be that situation. If the database is being used by more than one system - ecommerce where orders are coming in via web and via call center - changing multiple systems to accomodate a change in data may not be as effective as using a trigger.
I tend to agree with you, and in situations where the entire system is in one layer - one web server, for example - triggers might not make much sense. But if anything else is ever updating that database besides your application, having triggers available can be very handy.
I completely understand that row-level locking isn't a mandatory item - I'm not reading the Oracle marketing either!:)
What struck me about this was that I worked at a company in '98 using SQL Server 6.5 for most work. I was asking if we could use MySQL for some projects - we were doing unix/perl and NT/ASP. The NT devs all got to use SQL 6.5, and the Perl guys were only allowed to use flat files, no SSI on Apache, etc. Gave the NT devs an upper hand in being able to be more productive.
When I'd asked about using mysql, I got a couple nasty emails back from people who stated it 'sucked' because it 'didn't have row level locking'. One of them had just come back from a preview of SQL Server 7 which - surprise - had row level locking. The fact that we had a dozen LARGE clients (one doing > $500 million a year in ecommerce on our servers) running on a database server (MSSQL 6.5) without row-level locking seemed to escape pretty much everyone there but me. I'm *NOT* saying that MySQL would necessarily have been a good replacement for SQL6.5 at that time, but row-level locking was not the reason.:)
Yes, mostly mainframes, but I've no doubt that some industries were running "enterprise" apps 5 years ago on platforms that aren't as robust as MySQL5 is now. Yes, software has become more demanding in the past few years, but the fundamentals haven't changed. If you could run 'enterprise' solutions on SQL Server 6.5 (and I saw companies doing it - and gosh, they didn't even have row level locking!) surely some "enterprise" industries can use MySQL5 today.
... I'd like to give an example. Everybody has filled out a form with a request for vacation holiday to submit to your boss. You have to fill out the starting date, the end date, how many days you want to be gone, and who is responsible when you are gone. You first have to get the form, you download it, you print it, you fill it out. Then you mail it or you get it to your boss in some other way. Then the request gets granted, then somebody has to maintain the data base of how many holidays you have left, and so forth. It is slow and inefficient.
With web services and service-oriented architectures and X-forms, this process will be entirely different. You'll download the forms from your company's website, fill out the form, press submit button, the data will be sent to a web server which maintains the holidays left, and everything will get done automatically. It will tell you if you have enough days left, a notification will be sent to the person who has to approve the holiday application, and the whole process will be much smoother. This is how web flow will be done more and more over the next year or two. Having support for the end user this way will be a big deal, and will change how we think of collaboration with forms.
No offense to anyone involved here, but I worked at a company that was doing that over a year ago with Sharepoint/MSOffice. The backend technical details were probably slightly different than what they're talking about here, but lordy this is nothing revolutionary. The fact that OO is now offering a way to do it - maybe. The thing that bugs me is that reading things like this, I get the impression that people working on things like this (I don't mean vacation request systems, but many open source projects in general) is that features like this area touted out like they are something new or revolutionary. It indicates that they're probably not keeping up with with other vendors/platforms are doing. I wish I could put this in to words better, but I don't have any more time right now.:)
I'm a bit biased because I'm directly involved in it, but http://webdevradio.com/ aims to be a podcast covering web development topics. It's been a bit of a mixed bag so far, primarily because I'm learning the production part as I go along, but I aim to continue to make it better, both with experience and with feedback from people (hopefully) like yourself. We've recently joined the techpodcasts.com community, and I can say that there's a good selection of tech development podcasts there too. http://polymorphicpodcast.com/ and http://softwareas.com/ spring to mind (not sure if softwareas.com is a podcast or not, I just follow the site).
It's gotta be firefox/mozilla *plus* some setting or other program on your computer??? I've used firefox/mozilla since pre 1.0 days as well, on laptops and desktops under RedHat, Fedora, Ubuntu, Mandrake, Windows 2000, Windows 98, Windows XP and have never experienced that behaviour. I'm certainly not saying it's not happening to you, but there's more to it that just 'Firefox' itself, otherwise I think I'd have seen it. That's also not to say FF/Moz don't have bugs, or even bugs that crash the system, but I've never experienced that, and I've not met others who mentioned that bug either.
If done well, I'd find this useful. Using a 'history' as it exists now in firefox/ie isn't terribly helpful. Giving me a quick search through previous items, thumbnails of what pages looked like when I went there, etc. I'm not new in thinking of these - I know I've read or discussed these and other ideas with other people, but have yet to see them in any current modern browsers. Am I missing something obvious?
I just had a listener post his experieneces with web-based project management, and basecamp was among those he tried, with not a very good opinion of it. Based on the few remarks about basecamp, I'm rather surprised that a company so apparently publicly devoted to 'usability' overlooked such basic things.
There are two other major chains I know of using Linux in their POS - burlington coat factory (I think most people knew about that) and Valvoline oil change places. I noticed the Valvoline place I went to last year using some console app, but was just booting up and he logged in to a RedHat 6.2 system. I'm sure there are others - I don't often bother to look, but it's nice to see all the same. Who knows of others openly using Linux as POS?
Likewise the person who makes a "decision" without having any knowledge of more than one option - in other words, an uninformed decision - is often doing great harm to the organization they saddle with that decision.
I thought that was good, and tried to explain why auto-responders range between useless and evil, but had no success.
Perhaps because they're not 'useless' or 'evil' but actually 'useful' in some situations, like in business environments when people need to let other people know they're away and who to contact in the meantime?
I don't know your whole situation, but this is the sort of anecdote which gives the open source push a bad name. Yes, it has good names too, and lots of positive press, but these sorts of stories scare people away from trying OS too.
Business has fully functional setup. Someone replaces it with something less powerful from the end user's perspective. Replacer tries to explain why getting rid of functionality people relied on is 'better'.
???
There's usually a 'proft!' at the end somewhere, but I'm not sure where it would fit in in this case.
At least the number of paid vacation days. Doesn't sound like this company is all that special compared to any other company based in Europe. While I agree that it's nice that someone buys you ice cream once in a while, that's not really the be all and end all of life.:)
Believe it or not, there are times when people install linux and do not go out and buy new hardware specifically just for that purpose. ESPECIALLY considering all the hype from ubuntu-fans about how it "JUST WORKS" (often just like that, in all caps). I understand the need to use supported hardware, really. My bigger point about this is the "JUST WORKS" attitude, when in reality, it doesn't magically do anything more than any other distribution. If the drivers are in linux, they'll be available on pretty much every distribution within a reasonable amount of time. If the drivers are not available, all the fanboy shouting about something "JUST WORK"ing doesn't change the fact that the driver don't exist for Linux. Pretty straightforward. To repeat myself just a bit more, much of the excitement over Ubuntu comes from people who came from the LFS/gentoo/early debian world.
I've heard this comment for a year or so now, and finally tried Ubunutu two weeks ago. I'm coming from years of monstly Mandrake, but have recently tried Xandros, Yoper and a couple other distributions, and have installed Suse and RedHat as recently as 18 months ago.
Ubunut does not 'just work' any more or less than other distros. I would actually say it works *less* than some other distros at certain things. I've known a few people that haved *raved* about Ubuntu "just working", and I could NOT understand what they were talking about. I realized at least one of them came from the gentoo/lfs world where getting a system running is days of work, so in comparison, yes, it's great.
A few things Ubuntu didn't have which other distros had on the same hardware:
1. Automatic mounting of available Windows partitions Mandrake (and I think Xandros and Yoper, can't remember others) would make/mnt/windows partitions and automount them for me. Ubuntu doesn't do that, at least not out of the box.
2. Auto detect network printers. My wife has an inkjet shared on her eMac on the wireless network. Xandros (and I think the LE2005 Mandrake) auto-setup that printer and made it available via CUPS out of the box. No way of doing that in Ubuntu.
3. My wireless card wasn't detected. Doesn't matter what distro, it doesn't work out of the box - I need ndiswrapper and custom setup. Not bitching about that, but Ubuntu didn't magically make it happen.
A slight bitch about apt-get here too - it won't inform me of partially matching package names. In urpmi, if I run "urpmi ndis" it'll come back with a list of package names which match 'ndis' if there's more than one. Debian/Ubuntu, I have to use a separate command to search the 'cache', which is just frustrating. Again, for someone coming from LFS background, yeah, Ubuntu is a breath of fresh air. But there are distros that 'just work' out of the box - providing a much more robust environment - more than Ubuntu.
Generally, you need to do that with software in general, regardless of 'OSS' or 'proprietary'. IME, you can do more complete testing with OSS than with proprietary.
I've not seen browsers support it *natively*, but only at the 'here's some javascript code'. Mozilla *seemed* to have it some time ago in the lower part of the browser window frame, but it wasn't labelled, and doesn't seem to be in firefox when I upload files down.
I've had the same issues with open source projects *AND* even closed source products that were a 'business'. I was at a company which spent 5 figures on a time tracking system which was supposed to 'integrate' with MS Project. *After* purchasing, we found it didn't do what we were told it did. Caveat emptor, etc. but what do you do? It had 'documentation', a 'support' number with people answering the phones, all the requisites of what people consider necessary for a 'business', but the product was broken for our needs. We were *lied* to, flat out, but had no recourse short of legal action. Should we have pursued that? Possibly, but that is more money and time pursuing something which has an unsure outcome.
Yes, there are more bad/unprofessional OSS projects out there than good, but it seems to be an equal problem for software in general, not something which only affects OSS.
I'm sometimes concerned by some of the issues that were brought up, but then go back to thinking that these problems generally aren't solely the province of 'open source' but software in general.
Conceptual integrity We only have to look at the history of the electronic computer to see that the greatest advances in technology have been made by brilliant, strong-willed individuals, usually supported by a small team of dedicated engineers - not community-based projects.
Some of the best open source project (most, really) tend to be started and grown by a single person or a very small group of people. After a critical mass is reached, sometimes things open up to a larger community of contributors, but the projects are already fairly well established. Compare PHP and Python - perhaps not the best examples, but close to mind right now. Python was/is primarily done by one person, and PHP seems now to be more 'community' driven, and the results are that PHP tends to have more problems with moving forward (witness the recent 4.4/5.0.5 references-changed-behaviour issue). I don't see these types of problems happening in projects with one figurehead - at least not as much.
Innovation Yes, many open source projects are copies of 'closed source' software, but many closed source offerings are copies of other closed source offerings as well, all trying to address perceived needs in a slightly different way. I would say that it frustrates me that there's many more new ideas that could be implemented in mozilla or konqueror, for example, which aren't, and probably won't be until MS or Apple does them first, then there'll be a quick copy in the open source world. File upload progress bar is the first which comes to mind, and it'll be frustrating when MS comes out with it first (whenever that is) and watch others catch up (the built in WYSIWYG HTML editor in IE was another one).
All in all, 'open source' is at heart a method of software development, and has pros and cons. Most of the things that were mentioned aren't only an issue for open source projects. I'm working at a company which has paid money for a commercial product (accounting software and ecommerce addon) and things don't work. It's been two months and things still don't work right. We've paid money, had multiple vendors out on site, been on support lines, and they can't get it to work as it's supposed to. We're one of their first customers trying to use the software this way (I think) so this is a learning curve for them, and I've seen this happen dozens of times over the years. Why people think this is 'more acceptable' than having in-house developers working with free software, simply because you've 'paid' for something, is still a mystery to me. Downtime/lost productivity is not something you can get back, even if you get a refund of your purchase price.
Something I just blogged about (mostly just to make sure I didn't forget it!) was an idea for autoconverting docs via a mail system.
Yahoo Mail already seems to do a bit of converting some MS Office docs into HTML for viewing in your browser. What I'm talking about is the next step: autoconvert between openoffice and ms office.
I send someone an.SXW or.ODT file via Yahoo Mail. Y! converts the file int a.DOC file, then sends it to the recipient. They edit, send back, and it automatically converts it back to a.SXW or.ODT file (whatever my preference is).
I know there would be a lot of bugs and things that wouldn't work right to start with, but leave it in beta for awhile (perhaps gmail should offer this then?). However, I think the long term good could outweigh the short term drawbacks. Yes, there's a privacy concern, but if you're really that concerned about the docs you shouldn't be using public mail systems in the first place, right?
Slashdot Mirror
While this approach would work in certain circumstances, there's still an assumption there that the original system was built to accomodate that sort of scenario. If it was, great. If it wasn't, then you're stuck.
1) Triggers are hidden application logic that are very hard to debug, and are easily overlooked or forgotten by developers. Business logic (other than defensive logic like unique indexes, primary keys, foreign keys, not-null columns) does not belong in the database. They belong in the middle tier. They also make it much more difficult to move to another database.
I used to think as you write here, but have modified my thinking some in the past few years. I agree with your view on the 'hidden application logic', it might not always be that situation. If the database is being used by more than one system - ecommerce where orders are coming in via web and via call center - changing multiple systems to accomodate a change in data may not be as effective as using a trigger.
I tend to agree with you, and in situations where the entire system is in one layer - one web server, for example - triggers might not make much sense. But if anything else is ever updating that database besides your application, having triggers available can be very handy.
I completely understand that row-level locking isn't a mandatory item - I'm not reading the Oracle marketing either! :)
:)
What struck me about this was that I worked at a company in '98 using SQL Server 6.5 for most work. I was asking if we could use MySQL for some projects - we were doing unix/perl and NT/ASP. The NT devs all got to use SQL 6.5, and the Perl guys were only allowed to use flat files, no SSI on Apache, etc. Gave the NT devs an upper hand in being able to be more productive.
When I'd asked about using mysql, I got a couple nasty emails back from people who stated it 'sucked' because it 'didn't have row level locking'. One of them had just come back from a preview of SQL Server 7 which - surprise - had row level locking. The fact that we had a dozen LARGE clients (one doing > $500 million a year in ecommerce on our servers) running on a database server (MSSQL 6.5) without row-level locking seemed to escape pretty much everyone there but me. I'm *NOT* saying that MySQL would necessarily have been a good replacement for SQL6.5 at that time, but row-level locking was not the reason.
MySQL still has non-standard syntax and semantics
What 'non-standard' syntax are you talking about? Something like postgres' allowing LIMIT support in SQL statements? What 'standard' does that follow?
Yes, mostly mainframes, but I've no doubt that some industries were running
"enterprise" apps 5 years ago on platforms that aren't as robust as MySQL5 is now. Yes, software has become more demanding in the past few years, but the fundamentals haven't changed. If you could run 'enterprise' solutions on SQL Server 6.5 (and I saw companies doing it - and gosh, they didn't even have row level locking!) surely some "enterprise" industries can use MySQL5 today.
... I'd like to give an example. Everybody has filled out a form with a request for vacation holiday to submit to your boss. You have to fill out the starting date, the end date, how many days you want to be gone, and who is responsible when you are gone. You first have to get the form, you download it, you print it, you fill it out. Then you mail it or you get it to your boss in some other way. Then the request gets granted, then somebody has to maintain the data base of how many holidays you have left, and so forth. It is slow and inefficient.
:)
With web services and service-oriented architectures and X-forms, this process will be entirely different. You'll download the forms from your company's website, fill out the form, press submit button, the data will be sent to a web server which maintains the holidays left, and everything will get done automatically. It will tell you if you have enough days left, a notification will be sent to the person who has to approve the holiday application, and the whole process will be much smoother. This is how web flow will be done more and more over the next year or two. Having support for the end user this way will be a big deal, and will change how we think of collaboration with forms.
No offense to anyone involved here, but I worked at a company that was doing that over a year ago with Sharepoint/MSOffice. The backend technical details were probably slightly different than what they're talking about here, but lordy this is nothing revolutionary. The fact that OO is now offering a way to do it - maybe. The thing that bugs me is that reading things like this, I get the impression that people working on things like this (I don't mean vacation request systems, but many open source projects in general) is that features like this area touted out like they are something new or revolutionary. It indicates that they're probably not keeping up with with other vendors/platforms are doing. I wish I could put this in to words better, but I don't have any more time right now.
I'm a bit biased because I'm directly involved in it, but http://webdevradio.com/ aims to be a podcast covering web development topics. It's been a bit of a mixed bag so far, primarily because I'm learning the production part as I go along, but I aim to continue to make it better, both with experience and with feedback from people (hopefully) like yourself. We've recently joined the techpodcasts.com community, and I can say that there's a good selection of tech development podcasts there too. http://polymorphicpodcast.com/ and http://softwareas.com/ spring to mind (not sure if softwareas.com is a podcast or not, I just follow the site).
It's gotta be firefox/mozilla *plus* some setting or other program on your computer??? I've used firefox/mozilla since pre 1.0 days as well, on laptops and desktops under RedHat, Fedora, Ubuntu, Mandrake, Windows 2000, Windows 98, Windows XP and have never experienced that behaviour. I'm certainly not saying it's not happening to you, but there's more to it that just 'Firefox' itself, otherwise I think I'd have seen it. That's also not to say FF/Moz don't have bugs, or even bugs that crash the system, but I've never experienced that, and I've not met others who mentioned that bug either.
If done well, I'd find this useful. Using a 'history' as it exists now in firefox/ie isn't terribly helpful. Giving me a quick search through previous items, thumbnails of what pages looked like when I went there, etc. I'm not new in thinking of these - I know I've read or discussed these and other ideas with other people, but have yet to see them in any current modern browsers. Am I missing something obvious?
I'm interested in talking to you more about this. Email mgkimsal at gmail.com with some more thoughts. Thanks.
I just had a listener post his experieneces with web-based project management, and basecamp was among those he tried, with not a very good opinion of it. Based on the few remarks about basecamp, I'm rather surprised that a company so apparently publicly devoted to 'usability' overlooked such basic things.
http://fireboxstudios.com/news/newapp
There are two other major chains I know of using Linux in their POS - burlington coat factory (I think most people knew about that) and Valvoline oil change places. I noticed the Valvoline place I went to last year using some console app, but was just booting up and he logged in to a RedHat 6.2 system. I'm sure there are others - I don't often bother to look, but it's nice to see all the same. Who knows of others openly using Linux as POS?
Likewise the person who makes a "decision" without having any knowledge of more than one option - in other words, an uninformed decision - is often doing great harm to the organization they saddle with that decision.
Could you post it on a server someplace and give us the URL to look at it? It would be useful to see how other people address this issue. Thanks.
I thought that was good, and tried to explain why auto-responders range between useless and evil, but had no success.
Perhaps because they're not 'useless' or 'evil' but actually 'useful' in some situations, like in business environments when people need to let other people know they're away and who to contact in the meantime?
I don't know your whole situation, but this is the sort of anecdote which gives the open source push a bad name. Yes, it has good names too, and lots of positive press, but these sorts of stories scare people away from trying OS too.
Business has fully functional setup.
Someone replaces it with something less powerful from the end user's perspective.
Replacer tries to explain why getting rid of functionality people relied on is 'better'.
???
There's usually a 'proft!' at the end somewhere, but I'm not sure where it would fit in in this case.
Why do new words have to be invented for something, especially when they are just the lazy contraction of existing words that work perfectly well?
So they can be made into verbs. Try these:
"I'm going to post something on my development log"
vs
"I'm going to blog"
Same meaning, but catchier.
That was about as insightful as it gets...
At least the number of paid vacation days. Doesn't sound like this company is all that special compared to any other company based in Europe. While I agree that it's nice that someone buys you ice cream once in a while, that's not really the be all and end all of life. :)
Believe it or not, there are times when people install linux and do not go out and buy new hardware specifically just for that purpose. ESPECIALLY considering all the hype from ubuntu-fans about how it "JUST WORKS" (often just like that, in all caps). I understand the need to use supported hardware, really. My bigger point about this is the "JUST WORKS" attitude, when in reality, it doesn't magically do anything more than any other distribution. If the drivers are in linux, they'll be available on pretty much every distribution within a reasonable amount of time. If the drivers are not available, all the fanboy shouting about something "JUST WORK"ing doesn't change the fact that the driver don't exist for Linux. Pretty straightforward. To repeat myself just a bit more, much of the excitement over Ubuntu comes from people who came from the LFS/gentoo/early debian world.
I've heard this comment for a year or so now, and finally tried Ubunutu two weeks ago. I'm coming from years of monstly Mandrake, but have recently tried Xandros, Yoper and a couple other distributions, and have installed Suse and RedHat as recently as 18 months ago.
/mnt/windows partitions and automount them for me. Ubuntu doesn't do that, at least not out of the box.
Ubunut does not 'just work' any more or less than other distros. I would actually say it works *less* than some other distros at certain things. I've known a few people that haved *raved* about Ubuntu "just working", and I could NOT understand what they were talking about. I realized at least one of them came from the gentoo/lfs world where getting a system running is days of work, so in comparison, yes, it's great.
A few things Ubuntu didn't have which other distros had on the same hardware:
1. Automatic mounting of available Windows partitions
Mandrake (and I think Xandros and Yoper, can't remember others) would make
2. Auto detect network printers. My wife has an inkjet shared on her eMac on the wireless network. Xandros (and I think the LE2005 Mandrake) auto-setup that printer and made it available via CUPS out of the box. No way of doing that in Ubuntu.
3. My wireless card wasn't detected. Doesn't matter what distro, it doesn't work out of the box - I need ndiswrapper and custom setup. Not bitching about that, but Ubuntu didn't magically make it happen.
A slight bitch about apt-get here too - it won't inform me of partially matching package names. In urpmi, if I run "urpmi ndis" it'll come back with a list of package names which match 'ndis' if there's more than one. Debian/Ubuntu, I have to use a separate command to search the 'cache', which is just frustrating. Again, for someone coming from LFS background, yeah, Ubuntu is a breath of fresh air. But there are distros that 'just work' out of the box - providing a much more robust environment - more than Ubuntu.
Generally, you need to do that with software in general, regardless of 'OSS' or 'proprietary'. IME, you can do more complete testing with OSS than with proprietary.
I've not seen browsers support it *natively*, but only at the 'here's some javascript code'. Mozilla *seemed* to have it some time ago in the lower part of the browser window frame, but it wasn't labelled, and doesn't seem to be in firefox when I upload files down.
I've had the same issues with open source projects *AND* even closed source products that were a 'business'. I was at a company which spent 5 figures on a time tracking system which was supposed to 'integrate' with MS Project. *After* purchasing, we found it didn't do what we were told it did. Caveat emptor, etc. but what do you do? It had 'documentation', a 'support' number with people answering the phones, all the requisites of what people consider necessary for a 'business', but the product was broken for our needs. We were *lied* to, flat out, but had no recourse short of legal action. Should we have pursued that? Possibly, but that is more money and time pursuing something which has an unsure outcome.
Yes, there are more bad/unprofessional OSS projects out there than good, but it seems to be an equal problem for software in general, not something which only affects OSS.
I'm sometimes concerned by some of the issues that were brought up, but then go back to thinking that these problems generally aren't solely the province of 'open source' but software in general.
Conceptual integrity
We only have to look at the history of the electronic computer to see that the greatest advances in technology have been made by brilliant, strong-willed individuals, usually supported by a small team of dedicated engineers - not community-based projects.
Some of the best open source project (most, really) tend to be started and grown by a single person or a very small group of people. After a critical mass is reached, sometimes things open up to a larger community of contributors, but the projects are already fairly well established. Compare PHP and Python - perhaps not the best examples, but close to mind right now. Python was/is primarily done by one person, and PHP seems now to be more 'community' driven, and the results are that PHP tends to have more problems with moving forward (witness the recent 4.4/5.0.5 references-changed-behaviour issue). I don't see these types of problems happening in projects with one figurehead - at least not as much.
Innovation
Yes, many open source projects are copies of 'closed source' software, but many closed source offerings are copies of other closed source offerings as well, all trying to address perceived needs in a slightly different way. I would say that it frustrates me that there's many more new ideas that could be implemented in mozilla or konqueror, for example, which aren't, and probably won't be until MS or Apple does them first, then there'll be a quick copy in the open source world. File upload progress bar is the first which comes to mind, and it'll be frustrating when MS comes out with it first (whenever that is) and watch others catch up (the built in WYSIWYG HTML editor in IE was another one).
All in all, 'open source' is at heart a method of software development, and has pros and cons. Most of the things that were mentioned aren't only an issue for open source projects. I'm working at a company which has paid money for a commercial product (accounting software and ecommerce addon) and things don't work. It's been two months and things still don't work right. We've paid money, had multiple vendors out on site, been on support lines, and they can't get it to work as it's supposed to. We're one of their first customers trying to use the software this way (I think) so this is a learning curve for them, and I've seen this happen dozens of times over the years. Why people think this is 'more acceptable' than having in-house developers working with free software, simply because you've 'paid' for something, is still a mystery to me. Downtime/lost productivity is not something you can get back, even if you get a refund of your purchase price.
Something I just blogged about (mostly just to make sure I didn't forget it!) was an idea for autoconverting docs via a mail system.
.SXW or .ODT file via Yahoo Mail. Y! converts the file int a .DOC file, then sends it to the recipient. They edit, send back, and it automatically converts it back to a .SXW or .ODT file (whatever my preference is).
Yahoo Mail already seems to do a bit of converting some MS Office docs into HTML for viewing in your browser. What I'm talking about is the next step: autoconvert between openoffice and ms office.
I send someone an
I know there would be a lot of bugs and things that wouldn't work right to start with, but leave it in beta for awhile (perhaps gmail should offer this then?). However, I think the long term good could outweigh the short term drawbacks. Yes, there's a privacy concern, but if you're really that concerned about the docs you shouldn't be using public mail systems in the first place, right?