The Joel on Software Discussion Group (CLOSED)

A place to discuss Joel on Software. Now closed.

This community works best when people use their real names. Please register for a free account.

Other Groups:
Joel on Software
Business of Software
Design of Software (CLOSED)
.NET Questions (CLOSED)
TechInterview.org
CityDesk
FogBugz
Fog Creek Copilot


The Old Forum


Your hosts:
Albert D. Kallal
Li-Fan Chen
Stephen Jones

.NET with no linker.. yes but..

We've discussed that the run time is big and a burden...
BUT it seems apps built are quite frugal in size.. for example:

http://www.highlydeveloped.net/download.aspx

*Incidentally its a link to a free/shareware MS Project file viewer, pretty decent*

Agreed, that the app example is pretty lightweight but still..

So, once you have .net framework installed, isnt it a huge advantage for everybody from developers to end users because of the small download size?
FlashGord
Wednesday, March 23, 2005
 
 
another small sized .NET app, with decent amount of functionality:

http://www.joher.com/dotirc_download.shtml
FlashGord
Wednesday, March 23, 2005
 
 
> BUT it seems apps built are quite frugal in size.

Right, but, in this age of 160GB hard-disks, the whole issue is not that a 10KB EXE requires a 65MB run-time, but rather that most people still don't have that run-time, most of them don't have ADSL, and it's hell to get it from here to there.
Fred
Wednesday, March 23, 2005
 
 
Ironically, these exact same arguments were used in the past:

"Nobody will use Windows to run games! DOS is the lowest common denominator!"

"Nobody has 32-bit Windows yet! Our apps must remain 16-bit!"

In truth, .NET will get there eventually, just like today we don't make games in DOS any more, and all the Windows apps are 32-bit. Like these transitions, it will take time. Like these transitions, if you bet on the older technology, you will eventually lose.
Brad Wilson Send private email
Wednesday, March 23, 2005
 
 
What's so hard about checking for it during the installation and if they don't have it, install it for them? Just make sure you let them know during the install so if they should happen to see it they will know well enough to leave it installed. Seems blown out of proportion.
JG
Wednesday, March 23, 2005
 
 
Well if you want to demonstrate a .NET application to someone if there isn't the run time around its going to take a while to go get it, the vast majority of which isn't going to be required.

The equivalence would be true if you'd had to download and install Windows to do the same thing, but you didn't, it was always an OEM product.
Simon Lucy Send private email
Wednesday, March 23, 2005
 
 
The setup installer with VS.NET can be set to include the runtime (not good as an unnecessary fat download will be unwelcome), or to prompt with a download link to MS if the runtime is not installed (much better).

As for demonstrating to somebody - what's wrong with carrying the redist around on a USB key if you're going to be physically present? IF you're not physically present then how do you do a remote demo over a modem anyway?

Rich
poorHouse
Wednesday, March 23, 2005
 
 
And when those people are trying out a trial edition?

I know that I've not downloaded some shareware/trialware app because it wasn't a small download..
i like i
Wednesday, March 23, 2005
 
 
I'm sorry. I was thinking in terms of larger applications. In the shareware/trialware app market where the application is smaller I can see how this could be a burden. But that's a simple call. Wait for Longhorn. Unfortunately no amount of wishing is going to shrink the ASP.NET runtime, just like it hasn't for Java. Just stick with what works best for your situation in the meantime and take solace in the fact that in the future you won't have to worry about this.
JG
Wednesday, March 23, 2005
 
 
>> Well if you want to demonstrate a .NET application to someone..

In that case you would be carrying the runtime everywhere?

Isnt it obivious that eventually the runtime will be everywhere? If sufficient .NET applications come up, sooner or later u will install it.. and that'll make life easier for all subsequent installs and running
doogieHowser
Wednesday, March 23, 2005
 
 
"In truth, .NET will get there eventually"

No, it will not, because .net is a MOVING TARGET.

MS is now saying ".net has a 50% penetration", but :

1. They have not evidence. (I asked for backup on that # and the MS Product manager said "marketing told us that". yeah, MS Marketing)

2. They're talking about .Net 1.0 penetration, while extolling the benefits of .Net 2.0

3. Microsoft has a long history (ala "Innovator's Solution") of continual improvements, so we'll see a new .Net version every year or so.
(And they better. VB.net hasn't even caught up with C# IDE features like Refactoring, much less with Delphi 2005 fetaures).


So, your choice is using .Net 1.0 or 1.1 with lots of bugs (like memory leaks when displaying xp themes!) or jump to .Net 2.0 and count on installing the .net runtime.



"What's so hard about checking for it during the installation and if they don't have it, install it for them?"

1.  They need to have ADMIN privledges to install .net
More and more corporate users do NOT have those privledges. (This is becoming very common in Australia, for example).
This will become MORE of a problem.

2.  They may not have the patience to sit for a couple of hours on a dialup line (or their connection may drop).  I have some users that can't even download 10 MB files b/c of poor connections. This will become less of a problem, I admin.

3.  They may also have to download other UPDATE files. Read Joe's horror story about the security updates needed for .net, including LOTS of rebooting.  I don't know if your customers will REMEMBER they were downloading/trying our software after 2 reboots. Many of them FORGET where they PUT the downloaded file. Remember, I'm talking about Joe-Average user here.


"Isnt it obivious that eventually the runtime will be everywhere?"

See above. .Net is a moving target.


BOTTOM LINE
.Net was designed to compete with Java. Well, I dismissed Java for the same reasons: runtime issues.

.Net is a vast improvement for SERVER apps. Not designed (or used much) for DESKTOP apps.

You CAN (literally) shoehorn it into a desktop app, but it's an effort --FOR YOUR CUSTOMERS.


TRIAL DOWNLOAD IS CRITICAL TIME IN SALES PROCESS
Remember, the most fragile step in the sales process is downloading and installing the trial. That's where all the work is done by the cutomer and they dont' see ANY benefit till it's done.  ANY SCREW UP during that time and they may punt, having only seen a PROBLEM, but not benefit.  That's WORSE than if they'd never tried the trial.
Mr. Analogy {uISV owner} Send private email
Wednesday, March 23, 2005
 
 
Yes, and as soon as Microsoft releases a version of their OS with .NET as a native framework, most of these objections will disappear.

Except for the performance issues, of course, but with the "hardware of tomorrow" maybe that won't be as big an issue, either.
AllanL5
Wednesday, March 23, 2005
 
 
How many win2k boxes are there out there, who might be your customer?  How many will there be when that OS with .NET built-in is released?

And will that OS+.NET be out-of-date real quick?  Will .NET.Next be out by then?
i like i
Wednesday, March 23, 2005
 
 
"as soon as Microsoft releases a version of their OS with .NET as a native framework, most of these objections will disappear."

No. No. No. 

.Net is a moving target.



Longhorn will come out with .Net 1.1 or 2.0. It will take a couple of years for significant penetration (>30%)

By that time, Microsoft will be "encouraging" developers to be on .Net 3.0.

Even vb.Net 2.0's ide is behind Delphi 2005.  It doesn't have significant refactoring, for example.

You can use .net, even with the .net runtime distribution issues. It's just an obstacle to be overcome.

But, make no mistake, this problem of the .net runtime will NOT be going away.  You will ALWAYS have a significant % of customers who DO NOT have the CURRENT .net runtime.


Now, you can alleviate a lot of those concerns with Remotesoft's linker ($500, well worth it). Or use Delphi <g>
Mr. Analogy {uISV owner} Send private email
Wednesday, March 23, 2005
 
 
Mr. Analogy, I'm glad to finally have found someone who understands the limitations on PC use (ergo, difficulty of installing stuff) imposed by big companies.
NetFreak Send private email
Wednesday, March 23, 2005
 
 
I don't know. It still looks to me to be an issue mainly for those developers that are in the anti .NET camp anyway. They just seem to use the runtime size as one more excuse.

Suppose someone told you: "5mb a piece? Those mp3 sharing things don't stand a chance. Who wants to download even a single CD when it is going to be a 50Mb download job? Not everybody has DSL or cable you know!"

As for "Remember, the most fragile step in the sales process is downloading and installing the trial.", I'm not buying. All people I know seem to download stuff based on the promise of what it will help them do. If it takes 100Mb, it takes 100Mb. No problem if it is something that promises to solve a problem at a pricepoint they are willing to pay. The install better be without too much manual interaction, and you better not install any unwanted "toolbars" or "plug-ins" etc.
The most fragile step in the sales process is the first 15 minutes after they start up the app for the first time, with the first 3 to 5 minutes being the most critical in those. Don't piss of the client, think nice experience and get a quick result there seems to me to be the road to success.
Just me (Sir to you) Send private email
Wednesday, March 23, 2005
 
 
> It still looks to me to be an issue mainly for those developers that are in the anti .NET camp anyway. They just seem to use the runtime size as one more excuse.

It depends what kind of customers you have. If most of them have so-so IT, .Net is just a pain to use as of know, especially since it hasn't stabilized yet. In those places, just moving from something smaller like VB5 to VB6 proves to be a big PITA, so moving to .Net today would require driving to every single location, and install it ourselves on every single PC. At this point, it's just not worth it, but maybe in a 3-5 years.

For those markets, it's a better option to just keep working on the Win32 version, and wait until .Net has stopped moving, and is available as standard on most computers.
Fred
Wednesday, March 23, 2005
 
 
"..Suppose someone told you: "5mb a piece? Those mp3 sharing things don't stand a chance. "

You've just proven my point.

Audio file sharing didn't take off until the MP3 standard became common.  (OK, and Napster sure helped too)

You could have traded music files 10 years ago, but they'd have been 10x as big.  So, that didn't flourish.

So, at least in that case, SIZE DOES MATTER.

And remember, it's not just a matter of file size, it's also the admin issue and the windows updates issue.


And there's also the problem of: if my app installs microsoft's .net runtime which breaks something else on their computer (say, the .net requires XP SP2 and THAT can break something)  then guess who is blamed for the problem? *ME*

These problems can be overcome. They just demonstrate that .net is not designed for shrinkwrapped desktop apps.

By all means, use .net.  Just understand that the .net runtime is not going to go away.
Mr. Analogy {uISV owner} Send private email
Wednesday, March 23, 2005
 
 
Short of raw audio, MP3 is hands down the WORST format for disk apace usage. For equvialent quality, you can use 1/2 the disk space with WMAs or AACs.

The fact that MP3 flourished is in direct conflict with a few of your points:

1. People won't download large things (yes, clearly they will)

2. People will choose the smallest thing possible (not necssarily; MP3, despite being overly large, is ubiquotous)

3. People won't download arbitrarily large things (clearly not true for this case; the .NET runtime is a ONE TIME download that is about the size of 4 songs on P2P networks)

The "moving target" argument is just plain stupid. It's the argument of a newb, I'm sorry to say, as it doesn't reflect reality and existing practices. Windows is a moving target, too, and yet we still manage to use it as developers.

Face it: you argue the way you do, because you like Delphi. That's fine. Nobody here is trying to argue you out of using Delphi. Just watch your surroundings, or else progress will roll straight over you (as it is wont to do).
Brad Wilson Send private email
Wednesday, March 23, 2005
 
 
I have used .net with my application. The download is around 80mb with MDAC, .Net, MSDE. So far, no one complained. Most applications I use nowdays require ADMIN previlege to install. I remember downloading large trial download (> 600mb) from IBM. Either websphere or visualage. If people have no problem downloading pirated movies or large amount of mp3s, I do not see a major concern with the download size.
RC
Wednesday, March 23, 2005
 
 
Brad,

I never argued that people choose the smallest download. I argued that large downloads (>30 MB) will slow down adoption.

MP3s are about 2 MB apiece.  Comparable WAV file is about 20 MB or so.  WAV file sharing never took off. MP3 file sharing did. Main difference is smaller file size. Ergo, smaller file sizes helped that take off.

I've been developing software for over 20 years. Professionally developing for over 10 years. I'm not a newbie.

Windows is a moving target. HOWEVER, Windows 95 was basically complete. .Net is not complete. It's riddled with bugs. Just the XP Theme feature which causes a memory leak. That's just ONE bug. And it's in 1.0 and (I think) 1.1 as well.

So... .NET is not "finished". We'll have to wait for a finished version. Until then it's a moving target.

Microsoft can't get away with releaseing a buggy version of Windows. (Even Windows 95 wasn't as buggy as .net) to the PUBLIC. But they can get away with with foisting buggy crap on us developer.

I am not a delphi programmer, although I'm learning. I'm a VB programmer. I know about as much about .net as I do about Delphi at this point, so this is NOT a religious argument.
Mr. Analogy {uISV owner} Send private email
Wednesday, March 23, 2005
 
 
Oh, BTW, I do like Delphi. But I don't dislike .net b/c I like Delphi. Quite the opposite: I turned to Delphi b/c of the all the issues with  .net.

I'm just tired of my development environment getting yanked around by Microsoft's strategies. .Net is a strategy for combating Java. Understandable. But it means that they've abandoned support for VB6.

And since most MS developers aren't even using .net, why should I ?

.Net is great for internal development.  Once you get it out to every desktop in your organization, you've solved most of the hurldes. Not so great for  shrinkwrap.
Mr. Analogy {uISV owner} Send private email
Wednesday, March 23, 2005
 
 
I've bitten the bullet and decided to go the .NET route with my ISV. Not everyone is going to agree with me, but I'll put forth my reasoning.

1. Good software takes 10 years.  I don't plan on spending those years porting it from VB or MFC to .Net to whatever's next.  I will spend those years adding features and polishing the application solely in .NET. (And yes, I do have blind faith that MS will support .NET for 10 years.)

2. I am more productive in C# than VB or MFC (I've never tried Delphi, though).

3. My target customers are business users, which are much more likely to have high speed internet connections.

4. There is a nice body of frameworks, application blocks, and code generation tools for .NET that will enable me to spend more time on design and less time on coding.

5. I am concerned about the try-before-they-buy users that don't want to download the .NET runtime just to run a trial version. So, my website will include a fully navigable screen shot tour of each application.

Note, this list isn't comprehensive. I put together a weighted pro / con matrix of my options to reach my decision.
Cowboy coder
Wednesday, March 23, 2005
 
 
If .net is what you already know very well, and/or you will greatly benefit from the libraries and you use something like the Remotesoft or thinstall linker, I think .NET is a fine way to go.

For *me*, we didn't need the libraries, I don't know .net any better than anything else, and the dramatic price increase this year of Thinstall (it would cost us $80k plus $20k or so EVERY YEAR to use thinstall for our 20 programs) was a bit frightening. I hate being that dependent on such a niche vendor.

My criticism of .net is more a frustration that MS's only answer to writing software for Windows forces ME to distribute THIER RUNTIME.

If .net were such an easy upgrade, and so valuable, why doesnt' microsoft make it a required upgrade?

Instead, they are expecting ME (and you) to update thier O/S.

(I do realize the the OS folks and the VS folks are in different divisions, but if the VS folks really cared about us SHRINKWRAP developers, they'd find a way to make .Net ubiquitous OR provide a linker.
Mr. Analogy {uISV owner} Send private email
Wednesday, March 23, 2005
 
 
I'd like to give a vote of confidence to .NET here.  I'm the lead developer for an ISV that distributes a vertical market Windows Forms app.  It's moderate in complexity, perhaps 200K lines-of-code.

It is in active use in approx. 120 companies, with perhaps 2000 total users.

We have had no problems with the framework installation (aside from requiring admin).  None.  We've not had a single problem with instability, although performance of grids and such in Windows Forms often underwhelms.

In short, we're thrilled.  Others should not be scared off by a few naysayers.
Bill Carlson
Wednesday, March 23, 2005
 
 
"Even Windows 95 wasn't as buggy as .net"
I do not think so. I had no major problems with .net ( I do not use themes). Just because you found one bug does not mean the whole framework is full of bugs. If the customers of your software found one or two bugs and he/she declares your software is full of bugs, what would you think?
I am not a laywer but I suspect the reason not to include .net framework with windows was because of the antitrust law suit at the time. If .net is bundled with windows, Sun can make a strong argument about Microsoft's intention to kill Java run time.
RC
Wednesday, March 23, 2005
 
 
This reminds me of the resistance to VB early-on in the early 90s. .Net is a similar technological advance, a "better" way of developing applications, leveraged by a black box of runtime code. Same kinds of arguments - slower, bloated runtime, a "work in progress" were the chief issues raised by developers.

What happened is that these issues were gradually worked out in VB, partly due to increases in machine speed, partly due to better versions of Windows. But VB never became that big for shrinkwrapped mass marketed applications.

If .Net/CLR code becomes "the" native code for Windows in the future, then .Net applications should eventually become just as well accepted by consumers and developers as Win32 is today. If .Net remains a wart grafted onto the side of Windows, then "real" native applications (Win32) will be the best bet for the forseeable future.

The one big difference between the eras is that VB was a substantial improvement over C++ and other tools for developing common desktop applications. Whereas .Net feels quite painful for this type of work compared to VB or Delphi, for instance...
Bored Bystander Send private email
Wednesday, March 23, 2005
 
 
I'll have to weigh in on the side of Mr. Analogy.  If you're developing a desktop tool with any sort of reasonably sized market, .NET isn't really an option.

Our company's most popular download is around 600K (MFC app).  It runs on everything from Win95 to WinXP/Win2003.  A little over 1.2 million downloads thus far.  A .NET app would be a non-starter in this environment.

Think Google toolbar, GetRight, WinZip, or a million other Windows desktop apps.  It shouldn't be the job of the software developer to figure out how to patch an OS to the right level so that the app will run.

That's why MFC (or Delphi, or VB, or...) is still the only choice for mainstream GUI desktop apps.

Granted, if you've got a specialized (i.e., vertical) business app and you can tightly control the deployment environment, .NET may be a feasible option.

But for every one of those, there are a thousand mainstream apps that can't really use C# and .NET.

Of course, it goes without saying that for MSFT _server_ apps, ASP.NET is a no-brainer.  And PHP is a no-brainer if multi-platform web apps are required.

:-)
directorblue Send private email
Wednesday, March 23, 2005
 
 
>> And PHP is a no-brainer if multi-platform web apps are required.

Where have I heard that before...
Bored Bystander Send private email
Wednesday, March 23, 2005
 
 
"So, once you have .net framework installed,..."

Which version of the framework? Eg Avaya Voicemail Pro ( http://www.avaya.com ) requires you to remove v1.0 in order that it can install v1.1? So if your app wants .NET v2 on a machine running the Avaya software, what do I do?
DotNotSure
Wednesday, March 23, 2005
 
 
> In truth, .NET will get there eventually, just like
> today  we don't make games in DOS any more, and all the
> Windows apps are 32-bit. Like these transitions, it will
> take time. Like these transitions, if you bet on the
> older technology, you will eventually lose.

Not necessarily. 

In 1989, OS/2 was the hot new technology, and Windows was a kludgey hack on top of DOS.  People who bet on OS/2 lost big.  People who bet on Windows won big.

In 1992, WebTV and the Information Superhighway were the hot new technologies, and the WWW and Internet were crummy accomodations to the reality of dialup lines.  I bet you're reading this on a web browser and not a WebTV.

In 1996, CORBA and DCOM were the big distributed-object technologies, and the only reason someone would even *think* of SOAP or XML-RPC would be to leverage the large installed base of HTTP-capable software.  Are WebServices based on DCOM, or SOAP?

It's pretty hard to predict which technologies will become dominant.  You can point to Windows and 32-bit programs, but both of those had more technically complicated, more overhyped big siblings that have since been forgotten (OS/2 and RISC, respectively).  But since they *have* been forgotten, we don't even think about how else the future could've gone.

I'm lucky, I don't have a company to bet on one technology or another.  But if I did, I'd bet on what my users want *now*, not the promised big technology that's 2-3 years down the line.  At least then you have a war chest and happy customer base to quickly switch over to the new technology.  Bet everything too soon (in areas where it matters - Intranet software often has little or no penalty for using the .NET betas) and you can easily end up with nothing.
Jonathan Tang Send private email
Wednesday, March 23, 2005
 
 
"..If .Net/CLR code becomes "the" native code for Windows .....Net applications should eventually become ...accepted by consumers and developers

... If .Net remains a wart grafted onto... Windows, then "real" native applications (Win32) will be the best bet for the forseeable future.
.."

Do you think Microsoft is going to rewrite OFFICE anytime soon?

As long as office is Win 32, Windows will have fine support for Win 32. 

When do you think Microsoft is going to rewrite Win 32 as .net?

It took time about 5 years (assuming Longhorn is out on the newly revised schedule) to come out with Longhorn.

How long do you think it'll take to rewrite Windows in .net?

Not going to happen anytime soon.


Look fellows, I'm not saying .net isn't usable, I'm saying that the main problems with it (.net runtime) are NOT GOING AWAY for a very very long time, probably a decade. Runtime version is a MOVING TAEGET.

The pro .net folks like to say "these problems will go way". They won't.

BTW, the VB3/6 problems didnt' go away either.  The download size back then had (IMHO) more to do with the diskette restriction (and that DID go away with CDs) but also with DLL hell. DLL hell got WORSE not better, with ever version of VB.

And I loved vb.  But, here I am, 10 years later, with 16+ programs in vb3, with no upgrade path from Microsoft. (Don't EVEN start in on the VB3->4->6->.net converters. Tried 'em. They suck.)

So... here's microsoft saying "please adopt this whole new platform. Sure, it has some warts, but TRUST US. We'll fix it in the future."

BTW, all those .net folks in my VB User Group back in 2000 tried to tell me "oh, .net ships with Windows XP", now we're hearing "ships with SP2". It does NOT ship automatically. I know from experience. *Upgrades* will ship, but it won't just install .net for you unless you ask it to.

BTW, to the argument that "ms leaves out .net in sp2 b/c of antitrust issues". Ummmm.... then why are they including it in LongHorn?  Is DOJ 2.0 compabitble with .net, whereas DOJ 2005 is not? <g> (DOJ=Dept of Justice - the antitrust cops)
Mr. Analogy {uISV owner} Send private email
Wednesday, March 23, 2005
 
 
Mr. Tang makes some good points.


I'd always bet on an SUSTAINING evolution instead of a new innovation.

Hard Drives
Talk back in '98 was that SCSI drives were the future.  But IDE became EIDE and we've got pretty fast, big, cheap non-scsi drives.

USB ports
USB 1.0 was great.  Then we had firewire which was faster, now we have USB 2.0 which is even faster than firewire. AND USB 2.0 has the benefit of more ubiquity.  More people have SOME sort of USB port. So my USB drive works FAST on my 2.0 port, and at least WORKS on about 80% of computers.

Oh, and BTW, Delphi can write 32 bit apps today, then you can resuse most of that code and convert to .net if you ever need to.

KNow what's funny?  No one is mentioning why .net is any better than, say, delphi, or even ..cough... Java. All I hear is "it has lots of libraries" and "one day, it'll be the future".
Mr. Analogy {uISV owner} Send private email
Wednesday, March 23, 2005
 
 
Mr. A, take my comments more as attempting to find historical precedent than an opinion that .Net will prevail. I basically agree with you. All I was saying was that the way *may* be smoothed for .Net applications to become universal.

But - given that .Net on the desktop does not solve one lingering "problem" as such, and makes developer's lives much harder in the transition - it has tremendous resistance to overcome.
Bored Bystander Send private email
Wednesday, March 23, 2005
 
 
... and I am known as a Delphi bigot in circles I hang in.
Bored Bystander Send private email
Wednesday, March 23, 2005
 
 
FYI
Microsoft has settled with DOJ. So they are free to go back to their old ways until the next law suit.
The world is evolving. Learn from Wordperfect.
Adapt or die.
RC
Wednesday, March 23, 2005
 
 
>>> And PHP is a no-brainer if multi-platform web apps are required.

>>Where have I heard that before...

Argh... don't tell me certain parties have appropriated my lines!  ;-)
directorblue Send private email
Wednesday, March 23, 2005
 
 
+1 for Bill Carson; Thats what I wanted to hear..

Also, excellent discussions by Mr.Analogy, Bored Bystander and the others...

Anybody else who has actual experience building .NET standalone applications? I personally feel that the sheer weight of MS on developing and promoting .NET is going to make it ubiquitous.

Whether you/we like it or not is a whole different question. Like, maybe you would rather be developing commercial applications for Linux or Mac.. but are you doing it?
FlashGord
Thursday, March 24, 2005
 
 
>> first 15 minutes are crucial

Some very fundamental points:

1. When a customer NEEDS a functionality, he makes an effort to understand and make your application work. He is MUCH more tolerant than when you are trying to make a sell. If your application will work, he will download or get the required .NET runtime and if you provide it on a CD or give it as a link on your site he will thank you.

2. When YOU are trying to sell him something he hasnt even heard of or doesnt need currently, its tough. Thats when he will complain a lot about getting the runtime.
FlashGord
Thursday, March 24, 2005
 
 
Its definetly a matter of effort vs benefit. From my experience the .NET runtime is insignificant because the time it takes getting it setup right is nowhere near what it would take to configure our hardware manually. In my case if it takes them even 3 hrs to get IIS and .NET setup with an Admin on one server it's better than 3 weeks manually setting up the hardware with several hundred commands. But this is a deploy once, use everywhere app so it fits the .NET model more closely as it stands right now.

I think the problem is more a misunderstanding of where .NET currently stands. Obvioiusly it's a moving target, I've never seen software that wasn't. And all development environments have bugs, but you find out ways to deal with them and eventually they just become constraints until fixes prove otherwise. We've transitioned from Delphi/C++ to C++/C#/.NET, not without pains, but then transitions rarely aren't.
JG
Thursday, March 24, 2005
 
 
How many people have ATI graphics cards? if they keep there drivers up todate they probably have .NET and Catalyst Control Center. Given your monthly driver dl was 20 megs anyway its not a big deal. Lots of people download pretty much anything that is on windows update. Lots of people are willing to download 10-20 megs for a toolbar. I think there are probably lots of people who have .NET without realising it.

I can foresee a time where mono comes as standard with linux distributions. And they seem to be making good progress with their WinForms implementation.

Really if you are selling to companies it doesn't matter. If you are selling tothose with high end pcs/net connections they'll already have it or won't mind dling it. If you are selling to the average Joe they probably won't notice, its probably easier to install than the latest version of real player/random spyware.

If you are selling to those with Pentium 2s and 56k connections then you are probably selling to the wrong market anyway.

As for it beening a moving target these is very little that isn't these days.
Asd
Thursday, March 24, 2005
 
 
I have read some interesting reasons to go to .net, such as anticipated market penetration, not wanting to rewrite code etc...

It begs the question "Why?"

We already have great tools for writing windows only applications.  Why go to a new one?

Everyone likes a nice stable set of code, and .net might or not provide a good base (history indicates it won't)

But if you want a stable base, why not use java?  Worried about performance?  Compile the code.  There is no special reason to use the JVM if you are in a windows only environment, and you are covered if another OS does emerge.

FYI - I code professionally both Java and VB.NET.  But for my personal stuff I always use Java.  I do .NET only when clients require it.
flash91
Friday, March 25, 2005
 
 

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics
 
Powered by FogBugz