|
chaosgeorge wrote: OMG you want that a .NET version that stills for 10 years?
He didn't say that. He said he wants an app he writes to work that long. Maybe you can try and do better than just insulting him this time around.
|
|
|
|
|
Im not insulting him is just pointing that he looks he's against new Microsoft technology and i think we should adapt to it, by the way a .net 1.1 app run on vista now and in windows 7 with xp virtualization, and anyway in any os you can run your .net 1.0 app if not natively with virtualization, however i dont thnink apps that are that old are very productive these days
|
|
|
|
|
chaosgeorge wrote: Im not insulting him
I'm willing to bet if you reread your post you'll find plenty of items that can easily be deemed a tad bit less than cordial.
|
|
|
|
|
chaosgeorge wrote: he looks he's against new Microsoft technology and i think we should adapt to it
Ok, there lies our point of discussion: We disagree.
chaosgeorge wrote: by the way a .net 1.1 app run on vista now and in windows 7 with xp virtualization
Translation: Using Microsoft code today natively on their OS means that you can bet you will only be able to still use the application in a virtual machine in years to come when microsoft changes the rules again, making your already slow .Net software run even more slowly in a virtual machine.
chaosgeorge wrote: i dont thnink apps that are that old are very productive these days
Well, applications written today "Microsoft's" way when compared to "old skool apps" that do EXACTLY the same thing are 10 times larger and run slower - fact. Sloppy developers say "The machines are getting faster so it doesn't matter".... I argue that there is no benefit in speeding up hardware when the software developers do all they can to make new software run as slow as the old software did on on old hardware.
Here some real issues at hand for us developers that effect us all - personalities aside:
Issue 1: Rapid Application Development trades efficient run time for faster development cycles - PERIOD.
Issue 2: Investing in proprietary software from a company with a history of antiquating and abandoning their own creations only years after adoption is definately debatable.
Note: I've been a Microsoft Certified Consultant for years for more than one company - so I'm not out dated - in fact my feelings about these matters are from knowing the code inside and out. I know how many versions of .Net are out there as well as what is involved coding to them, using them, distributing them etc. What I am is tired of explaining to customers who spend tens or hundreds of thousands of dollars on a particular technology and finally getting their business truly "working" just to find out that all that code they wrote is on the Microsoft chopping block, and by the way... because you chose microsoft, all the microsoft compatible software you paid for... like Front Range Goldmine... Sales Logix and much more ... (games perhaps at home, video cards, etc) all need to be scraped because the new operating systems don't support it.
So... your opinion is "just get with the times".... where I'm moved by the fact that many of my own customers feel ripped off... because of the Microsoft "goo" they are stuck with - much of it I was their saying "yeah its good stuff".... and it was... Microsoft themselves have turned it into the past tense.
This is where I'm coming from.
Know way too many languages... master of none!
|
|
|
|
|
Have you used an hyper visor for virtualization? tell me if that's slow.
Well i think software is made to evolve to break RULES to change RULES and get them better that's the only way software technology is evolving and yes maybe SOME things goes just to broke, but that's the culture we have been lived for years: "if i have a software i want it to live forever and ever" even if you know it's costing thousands to mantain firstly because there are less and less people who is willing to mantain that kind of systems (and i count as one) and secondly because it uses old standards that no one longer uses, i think the culture of people who adopts software must change and say ok i have a software this year but i want in two years to be reingeenered or evolved so i can have more stable, better quality, better ROI software and overall MORE USEFUL software, and guess what today tools let that happen more and more softly dont tell me it's the same programming something in old school MFC c++ than programming it with today .NET or java technology for example, and of course having to mantain old software to be again mantained with old technology will cost more than if they had migrated before, dont you think? i think early adoption of new technology at the end costs a lot less than adopting it later.
JasonPSage wrote: What I am is tired of explaining to customers who spend tens or hundreds of thousands of dollars on a particular technology and finally getting their business truly "working" just to find out that all that code they wrote is on the Microsoft chopping block, and by the way... because you chose microsoft, all the microsoft compatible software you paid for... like Front Range Goldmine... Sales Logix and much more ... (games perhaps at home, video cards, etc) all need to be scraped because the new operating systems don't support it.
Ok you can't stop technology just because no one wants to break their apps, thats for hardware thats for microsoft and that's for all the other people investing in technology, and well OSes just have to adapt to change too, and that's in every OS my friend. And finally spending hundreds of thousand of dollars on a particular technology is just bad planning, you first need to test new technology before adopting it, and if you have adopted it you must have a plan to use it not just let it cook for you, and finally a plan to migrate it when the time comes, a well informed and organized organization won't lose with technology, they will win, and that happens with Microsoft adopters, Linux adopters, etc., and well Microsoft keeps going and going investing in technology not because of the people who loses with "choped" technology but because of the people who wins with their technology that are the most these days.
|
|
|
|
|
chaosgeorge wrote: Have you used an hyper visor for virtualization? tell me if that's slow.
Virtualization by definition is slow, but with fast enough hardware is great as you can "back up" entire OS's to a disk file - and restore them. As for your HyperV, run that same OS directly on the hardware: its faster - naturally. Forcing people to use virtualization is saying by the way, that new hardware you have won't be allowed to run your old app at full speed, as we only support it in a virtual machine. You can run anything you want if you throw enough hardware at it. Not everyone is willing to turn over all their server hardware everytime an OS is released. Staying current is one thing, providing there is enough income and ROI to fund it.
chaosgeorge wrote: Well i think software is made to evolve to break RULES to change RULES
Well, I've been innovatively coding for years and I couldn't agree more. That is not where our opinions stray. My problem is when companies like Microsoft literally force thier customers to abandon "what works" and antiquating what they have. If you are making new software, you don't have to break old stuff.
If you want to write new awesome things - go do it - just don't force feed it to everyone who might not want your new technology. New technology: Cool. People adopting new technology because they like it: Cool. Force feeding people new technology and acting like its the new salvation for mankind: Microsoft.
No one is trying to stop technology here or even suggesting it. Think of it this way: Architecture...like buildings and stuff. New technology brings us new materials to build with, new processes, faster means to create etc. THANKFULLY every time a new architect teams up with new material engineers and planning teams they do not destroy every building on the planet. That's what I'm saying Microsoft has been doing lately, and that is what I'm saying is ridiculous.
Know way too many languages... master of none!
|
|
|
|
|
JasonPSage wrote: Force feeding people new technology and acting like its the new salvation for mankind: Microsoft.
he lol well i don't think microsoft is acting like that, the times Microsoft had encapsulated everything are over and the times that the systems could not interconnect are over too and Microsoft is investing also on these technologies like everyone, and the future is that OSes will be less and less relevent as time goes by, and an enterprise who still thinks as the OS or any software as its one of the most importants values with no planning for the future is in danger
I think any good engineer could take the best of everything and integrate it as part of the whole, at the end also old buildings that didn't evolved are fully destroyed to build new ones  anyway i think people are a lot more free to do as we like than before and no one forces anything
|
|
|
|
|
chaosgeorge wrote: OMG what kind of systems do you code? lol i dont even remember what i was coding 10 years ago i guess it was basic or C, i think you are a frustrated out of date dude who could not catch up to latest technology
LOL, well you're right about one thing: I'm definately frustrated. I'm not outdated - in fact, if I was I wouldn't have the "tag line" I do: "Know way too many languages and master of none!"
The kind of systems I TRY to write are ones that stick around. Ten years ago I wrote a warehouse management system for one of the largest natural food food distributors in the United States on a HP-UX when a vendor specializing in such systems failed after receiving 3 million bucks. I wrote it in a month, reporting projects followed for another month or so to get everything needed for the Day to Day needs of people on the warehouse floor and for management.
You know what? They are still using that same system: works like Walmart's with the hand held scanners, the forklift scanners and wrist units from Symbol (for scanning bar codes of pickslots and products etc). It has matured over the years but its still the same system. that's the kind of systems I was writing 10 years ago, and I haven't stopped coding or staying current because I'm addicted to this stuff! I've been programming for 27 years and that my friend gives me a pretty good point of view over what technology is, does and what is innovative and what isn't.
In all fairness, Sharepoint and Microsoft CRM are decent packages; Server2008 is pretty decent as well. I particularly loved VB6 because the "package" was complete and it made true binaries unlike .Net (regardless of generation). Just In Time compilers do not protect intellectual property of developers unless they go through obfuscation steps to "conceal" their code.
Now for innovations - Microsoft's run of "Basic", Open Database Connectivity (ODBC), Various operating system roll outs prior to Vista and excluding Millenium, Outlook, Excel and Access and the Visual Basic for Applications tools, Microsoft Sidewider Joystick and various "Mice" ..well - awesome.
As for me being outdated, I disagree. New Technology is great however all to often the clammouring of new technology is just marketing folks putting proverbial lipstick on a pig and calling it new. I've been around long enough to know that today's "New" technology rarely is innovative. "Things" labeled as new technology often promote other people owning/managing your data (security risk), use more bandwidth than necessary (xml/soap as the interoperabilty cure all) and often it changes "processes" people know just enough to force them into learning how to do the same things thhey already do, differently which slows down productivity.
Here's an example of the kind of thing that frustrates me to no end: For as long as I can remember, in Microsoft products, the key combination CNTL-F pulls up a "Find" window where you can search for text and CNTL-H would call up a "Search-N-Replace" window: Wonderful. Now, another wonderful keypress I know and use without even needing to think about it, until recently, was using the F3 Key to "Find Next" instance of something I was searching for. This F3 key has been in multiple versions of Word, Excel, Access, Outlook, Powerpoint and many Microsoft programming environment IDE (Integrated Development Environment). This F3 "Find Next" keypress has been so well received that it can be found in all kinds of software these days regardless of operating system! But guess what... Microsoft doesn't use it anymore. I fire up Microsoft Office 2007, and it doesn't work the same way which means I have to work differently to do the same thing I've done for years without any appreciable benefit aside from inconveinence. Am I frustrated? Surely. Outdated? Just older and wiser.
I've programmed all kinds of systems including Microsoft CRM and Sharepoint and the API's are wordy, expensive from a bandwidth perspective, the interfaces are too "click happy" and users complain. You need people with serious experience in programming to beable to do anything more than add a custom field. You need to know a ton of technologies to understand how it all pieces together. Now I ask you friend, compared to Microsoft's "old" approach of having one script language that can talk to all it's products in the same manner... which is more intuitive? If this kind of observation and frustration is a sign I'm outdated, then so be it.
Know way to many languages... master of none!
|
|
|
|
|
JasonPSage wrote: labeled as new technology often promote other people owning/managing your data (security risk), use more bandwidth than necessary (xml/soap as the interoperabilty cure all) and often it changes "processes" people know just enough to force them into learning how to do the same things thhey already do, differently which slows down productivity.
Well about this you're right and well thanks to SP1 of .net 3.5 now we can use Data Services aka RESTFUL data or Resource Oriented Architecture, if we focus on that, yes its different way of do database stuff than it was done before (WCF, SOAP/XML) sure it is an innovation, of course ROA is not new but incorporating it .net and integrate it with linq well thats what's worth labeling "innovation through technology" making it easier, more elegant and more stable than with traditional WCF,SOAP,etc so my point is: is not just about making the same things in different ways, it's about achieving the same goal in less time and with better quality that's different and worth to invest.
So I conclude with this: the "new" approach of Microsoft is making every new technology a FOUNDATION, extensible technology to adapt to your needs, it's no longer that what you learned from VB6 is useless in vnext or .NET, since .NET era what you learned in .NET 1.0 applies the most for version 2.0, 3.0, 3.5 etc, that's why the major technologies of .NET have the sufix Foundation, and that's the same approach with Sharepoint and other technologies and that's something no other company has achieved, learning what's new is just a matter of weeks if you dedicate just part of your time, regards
|
|
|
|
|
Years ago my team and I put our truct in IBM and OS/2. We saw this Windows thing as a low quality FAD... then we moved our C++ development to cross platform tools and delivered for our OS/2 and our Windows customers - then all the cross-platform vendors either went out of business or decided to bleed us in licensing fees.
Then my company decided to go 100% COM/C++... in 2001!
I used to really hate Microsoft - all marketing, no quality! (this was from the DOS/Windows days). But IBM certainly was the wrong company to back.
Since then - Microsoft has done tons to improve productivity and quality. I can accomplish on my own what it would take a team of 4 or 5 people to produce in 1995. But, this comes at a price of technologyu shift. If you're in the old native C++ development mode - you're spednign way too much effort inventing the basics over and over.
Move on up to C# and ReSharper! Toss in some IoC - and lots of big I interfaces! Decouple that code and set yourself free! 
Dale Thompson
|
|
|
|
|
but it runs so slow...
Know way too many languages... master of none!
|
|
|
|
|
Jason: I feel your pain. I've written lots of different things over the last 25 years and what used to take 500 bytes in assembly now takes 500k (not counting the 20MB framework you ALSO need). The assembly used to be elegant, tight, every byte on the stack accounted for and ran like lightning even on the most modest of machines. The new stuff? It's a crap shoot. You no longer REALLY know what's going on since you only wrote two lines of code and the framework magically does the rest behind the scenes.
But is that really so bad? What used to take days or weeks I can now do in hours -- And much of it is less error prone! The whole concept of super high level languages and wrappers around wrappers that wrap yet something else may not be efficient, but it no longer HAS to be be in most cases. Hardware has gotten so cheap that, as much as I like optimizing code, it's often cheaper to just buy a bigger box. Don't get me wrong: There's a plethora of clueless application developers (versus programmers!) out there that take this concept to the extreme and just slap something together with total disregard for performance.
I agree with many of your points:
- RAD trades efficiency for rapid development cycles. Absolutely! But that's the point, isn't it?
- MS does release technologies at a pace that's impossible to keep up with. Their "let's see if this one sticks" approach drives me nuts! But stick to their core technologies and it's actually pretty solid. I've seen more backward compatibility out of MS than out of most companies.
As for the product at hand: Sharepoint...
We use it. It's a love/hate relationship. From a developer's perspective I think it's a pain in the neck. From an end-user standpoint: They "just get it". Show me anything in the same price range that has such a consistent UI (think users used to Windows/Word/Outlook/Etc.) and a similar feature set and I might be tempted to switch. I sure haven't seen it, and it's not because I haven't looked.
Peter
|
|
|
|
|
Peter: Well Said. I agree with EVERY word you said through and through.
On your point about Microsoft being pretty solid for "backward compat - I agree but as of late I keep finding this "solidarity" going south. Sharepoint? Good product but I'm ultimately switching my loyalties.
Note alot of my frustration comes from what I see customers having to go through more and more lately. I'll have you know - I have really loved working with Microsoft technologies for a long time, but lately they have been giving me more reasons to dislike them.
BTW - I'm a fan of RAD dev.. especially for GUI development and integration into existing foundations. I've written my own RAD systems that literally wrote tons of code. Shoot, I'm not about rewriting the wheel all the time, I'm a firm believer that a SOLID API lays the foundation so you can concentrate on the applications - where the API is a CONSTANT, applications are not. when the API changes so frequently (the calls and such) it's hard to keep up. Honestly, if the "guts" perform decent, I don't care about how they work. But I would be so much less frustrated if API calls stayed more constant and were added to. If this kind of API management was done properly... then applications could more easily be migrated, often without doing anything other than recompiling against the new API. If you told me you wrote an API that ran on ANY OS - and was always being added to... I'd love it. The problem is when calls you made before no longer work. for example... Microsoft did some good stuff in win32 api (and some bloopers) but mostly good. when they added a new feature, the call was usually named like "windowcreate_ex" for extended... and more than once the code behind the API functions was changed but the results were the same so you only needed a simple recompile.
If Microsoft has one BASE API that was endian independant and platform independant, performed decent, and compiled to binary on any platform (at least all their own OS's) I'd be pretty interested. If they didn't keep dropping support for old commands during its life and instead just changed the code underneath as improvements were added, and added new function calls for "recommended" new versions of calls for new software being written, I'd be very very impressed. Then we could get really really skilled at that foundation, then we could write REALLY awesome RAD software that wrote code for that platform so we get the job done faster, and then customers could keep building on their infrastructures rather than simply replacing "tried and true" tested systems for new stuff that hasn't been proven.
Know way too many languages... master of none!
|
|
|
|
|
JasonPSage wrote: If Microsoft has one BASE API that was endian independant and platform independant, performed decent, and compiled to binary on any platform
LOL! I think they did try that. Remember NT on Alpha? That worked well! At this point, the PC based platform is so ubiquitous that I don't think it's an issue anymore. You really want to run Windows on a Sun box?
Your comment on the Win32 API is spot on: It used to be rock solid. What worked on Win16 even worked on Win32 and it all remained compatible. Although they goofed up once in a while, stuff usually got added instead of changed. But the other side of that coin is that you have to have a huge amount of code in place to support those old functions in light of new hardware and new functionality. Net result: Things get slow and bloated. Isn't that where we started?
Also, some of the old functions just CANNOT be made secure. Look at something as simple as strcpy... How many functions do you know that just take a pointer and start writing there without a decent size check? And you know that stuff exists on any platform, not just Windows. With more an more developers that don't know their mallocs from their frees, it's a disaster waiting to happen.
At some point, you have to address this stuff. You have to get rid of some of your original (good or bad!) design decisions. You have to remove some of the bloat. And the market says you have to make it easier for less experienced programmers to write decent code in a reasonable timeframe. This gives you no choice but to drop some compatibility. This applies to the base API, but also to everything else.
They took this to the extreme when they released Vista and all hell broke loose. Now people (developers) are getting used to it and it's getting better. Windows 7 is the optimization of what they built in Vista (doing exactly what you said: Change the code behind the APIs, but not the APIs themselves) and in Windows 8 we'll all get it right.
The dotnet stuff addresses the last issues and they've actually done a pretty decent job. It's gone through some rapid changes in the first few versions, but I'd rather them do that early on.
And if you really want to run your "old" stuff, you know what? Run it on an old environment. We'll even give you a free virtual one. It's not entirely bad?
BTW: Change happens. The wipers from my old Dodge don't fit my new Bimmer either. And I thought the ones from the Dodge were fine; they removed the rain, didn't they?
Peter
|
|
|
|
|
I agree with you Peter. I'm not against new stuff - I'm for it.
I do disagree on the strcpy stuff though, because frankly that code isn't bloated - its lean and mean and in the right hands its plenty secure. Microsoft isn't the only one who knows a malloc from a free
I do like and embrace the idea of having options: strcpy "raw" as well as having newer more secure options with documentation that specifies which is recommended but allows developers to use their own judgement.
I'm good with change too. BTW - NT on Alpha was a decent approach and I was a true fan boy of NT and was for years. Server 2008 is pretty decent also.
--Jason
Know way too many languages... master of none!
|
|
|
|
|
well said friend you r quite right about MS release technologies so fast. List we just start using vs 2008 in past 6 month and now vs 2010 is coming 
|
|
|
|
|
Which is fine - 2008 and now 2010... the historical fact is - code you got working in 2008 might not recompile/work in 2010. Which makes it's so you aren't building on what you learn, you're too often scrapping what you had figured out.
I like to figure out new stuff... say I write a cool app that does XYZ... I should be free now to write a new application that does something different without being forced to constantly revisit and rework, retest the XYZ thing every "iteration" "product release" because someone else dictates as much.
I was an old turbo pascal developer years ago when Borland had a lot of awesome engineers writing their stuff. I left all that for Microsoft technology. Now, there is a language called freepascal which is pascal but it's completely object oriented like delphi, compiles on iphone, mac, linux, unix, dec alpha, nintendo and various embedded systems - and the API's have pretty much stayed backwards compatible and there even is a complete GUI IDE environment - where you write a gui once and it works on all the above OS's if they have a GUI system... even FreeBSD and OS/X for the people who really don't like change LOL. It even has a Delphi compatible mode where its like 99% dead on compatible with Delphi making delphi programs capable of compiling on other hardware and operating systems (when the code isn't to OS specific e.g. tons on win32 calls)
I wrote a web server with it that blows the doors off of IIS, apache and Lighttd - and unlike java's pcode - (java is not really compiled) - freepascal compiles to binary on every platform it supports. It's as fast and in many cases faster than microsoft's C++ stuff and runs circles around .net for speed.
I've used that language for years and it's funny - people still think pascal is a learner's language and they don't seem to know its object oriented. I've been heckled over my admiration for it often - but I laugh when my code runs faster and on more platforms than any other system I know (that does binary). Java does good for portable scripting, as does python, php and I think Ruby is honerably in the scripted/portable solution ranks now.
Ah well - the fact is - I'm an enthusiast and I love coding, I just like to build on what I've done and not have to start over each each year or two - especially because most decent software projects I write take a year or more to complete (the thinkstuff anyways).
--Jason
Know way too many languages... master of none!
|
|
|
|
|
|
JasonPSage wrote: Code I wrote 10 years ago on Unix still runs on Unix and Linux today. Code I wrote two years ago is already "outdated" in Microsoft Circles...
It's only a recent issue IMO, but yeah I totally agree. They're overdoing it. I don't see a real huge boost to productivity overall either compared to what I saw in the VB6 days.
Oh be careful with saying anything wrong about MS here. The kiddies are lurking. 
|
|
|
|
|
Well i am agree with you but, frankly i didn't like my code to be working after 10 or 15 years 
If our code works after that much time we will soon to be work less 
Think about this if your car work very good after 20 years you may not buy new one.
Every product should have there life span. Just my opinion 
|
|
|
|
|
The Microsoft ODBC technologies that have worked for years don't even work anymore on the new systems...
Really? I think you wrote this without thinking. It works for me.
Nuclear launch detected
|
|
|
|
|
So, help me on this one.
I have a legacy vb6 application that does all kinds of awesome things on ODBC databases. I've heard about and read about and seeked and search for MDAC8, as it's supposed to work on Vista. Have you had luck with this?
Further more, on the 64bit roll out of Vista, how do you proceed?
I'm not trying to be smart here - I haven't had any luck here and became frustrated.
There are literally tens of thousands of lines of tried and true code that allows me to copy databases from one to another like sharepoint but its easier. I can use it to monitor multiple databases remotely, launch programs and scripts when systems go down/connectivity lost, and likewise when it returns. I can monitor 250 databases with it... I can literally connect oto oracle, copy the entire datamodel to multiple target databases of different platforms, then export data models to excel, mysql, postgres, import and export to text: you name it... etc etc. It's a pretty awesome utility.... I'd hate to chuck it, rewriting it isn't really an option. It took on/off development over five years to get it where it is: its quite mature.
Know way too many languages... master of none!
|
|
|
|
|
Yes, ODBC 32-bit works normally on Windows Vista x64.
x64 applications works as well, only I tried only with SQL Server.
But everything works as normal.
I'm talking about C++ and ODBC API here. Not sure how VB6 is affected by MDAC.
I'm sure there is a MDAC 2.8 I think for 64-bit - I dont know if there is ODBC x64 support.
You can run application in 32-bit mode, even if is x64.
On Vista, I bet there's something related to UAC. You may try to read/write ODBC config files or so and you get all kinds of access denied or so errors, I don't know.
But this can be easily overcome by writing Vista specific code, or devising a CoCreateInstanceAsAdmin code sequences when UAC is needed. My 0.02.
Nuclear launch detected
|
|
|
|
|
Thanx for the response. I haven't much luck here - but I could try again.
Know way too many languages... master of none!
|
|
|
|
|
Everytime I INVEST time and energy into a Microsoft Product, they discontinue it just as I'm mastering it.
I'm mastering Windows API. (I try to do so every day since 1996). And that's not outdated.
Nuclear launch detected
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
After this I am going to code in binary.