Like I said, in some parts i'll call "bullship!" and in other you'd have to say outloud "yeah...my bad..." ...
Learning for Silverlight = Poor. Performance of Silverlight = Great (multi-threading, gpu/cpu access etc). You just probably needed more guidance but even as I type this I hear myself "Yeah sure, blame the developer for not knowing how to run the runtime...that's smart..." and slowly let my eyes fall to the ground with a mumbled "ok...fine...our bad.." Debugging of Silverlight = Meh, depends on what methodology/framework you're using to data snack between server & client. I've personally never once ran into that issue and i'll even lump Silverlight for mobile into that equation (XBOX Halo Waypoint for windows phone for example took a lot of engineering muscle to take large XML packets into a phone from the XBOX Live services team..which meant sucking it in, processing it inside the phone and then displaying it...and that was early bits of Silverlight on mobile..and still...debugging wasn't as bad as you outline...) .. that being said... i lost credibility (didn't have much to begin with) when I said "depends on the framework/methodology" ...which translates for me to "derp derp barnes, you're a jerk, derp derp" :) Productivity games i'll agree with. If you've never really gotten used to plugin development in terms of "RIA" (what we kids used to call it). It's a complete 180 in terms of behaviour traits in developers, as I noticed that even in Flash/Flex circles often developers would echo the fatal words "ahh screw this, i'm going back to my comfort blanket HTML/JS.." as at least the mediocrity is stabilized and time is slower for them to figure out state management and datagrids are solved with big fat TABLE tags (oh sorry, now its "div tags with CSS hacks") (Yes I roll my eyes at HTML/JS) :) Bah.. i think about the whole first years of Silverlight and i wish I had have pushed harder on a select few to stop being jerks and listen to the needs of a few other people to get funding so we could fix this crap and then I still come back to a conversation I had once with the leadership team around "Well...what happens if we do beat Flash..." to which one said "Well we'd break the glass, take the briefcase out, open the envelope and inside it would read - Congrats, now burn this plugin and focus on IE" So it was inevitable i guess :) (I think i'm up to Stage 5 of Silverlight Grief). --- Regards, Scott Barnes http://www.riagenic.com On Thu, Feb 13, 2014 at 11:39 AM, Tony Wright <tonyw...@gmail.com> wrote: > Alright I pay some of that. Within the corporate environment I was in for > the major project, we were also coding on 32 bit machines. We had > continuous memory overflow issues, which we raise to MS and were told it > wasn't going to be fixed because it was 32 bit issues. Together with the > crashes and pauses caused by switching to visual XAML view, development > became a major hastle. We eventually had to turn off visual view for > stability reasons. > > > > As for the async stuff. Yes, async is always going to be hard to debug, > but for some reason, it was even harder in Silverlight. Sure, if you get > the pattern right you shouldn't have any major problems overall, but having > to use fiddler to do simple diagnosis, a third party tool, was bizarre. > > > > In the end, when we looked at the stats, we found that developers were > heaps less productive with Silverlight than with web apps. It actually took > an embarrassing long time for people to get a Silverlight app debugged > relative to a web app. What took a week or two to do in web seemed to take > a month in Silverlight. We can talk all we want about the reasons for that, > and what gaps existed in developer Silverlight knowledge, but the reality > was that this became irrelevant with a significant team of developers. It > just took them far too long to learn and develop. > > > > > > > > *From:* ozdotnet-boun...@ozdotnet.com [mailto: > ozdotnet-boun...@ozdotnet.com] *On Behalf Of *Scott Barnes > *Sent:* Thursday, 13 February 2014 12:26 PM > > *To:* ozDotNet > *Subject:* Re: Migrating TFS > > > > Yeah i'm calling "Spacer Problem" on your issues described (ie the space > between Silverlight and you). Everything you've outlined can be distilled > into "We haven't yet figured out how deep linking as a concept works > because we're used to HTTP handling that burden for us" through to "Async > is hard stuff" > > Devs tweaked XAML because some idiot in the Visual Studio team decided to > let devs access XAML natively in the tooling, but forget to check off the > performance issues that come with Design vs XAML view. It was a conditioned > response to that problem and given XAML was really never meant to be a code > centric workflow it just baffles the mind at times as to who was actually > in charge of that mission and how they justified it to business reviews we > had. Blend was also a huge issue given they had zero time for Ux > stabilisation(s) and no real investment was given to that team to make it > User Friendly - even thought the target audiences were always that > "designer with dev experience" background(s) who are hyper sensitive to bad > Ux (What could possibly go wrong with this vision of the future). > > Silverlight also had a really crappy onboarding process, where we > basically walked up to the entire .NET community, kicked the crutch out > from underneath them and kept giving confused looks when they'd keep > falling over... that is the whole learning process for Silverlight was > spread throughout the web and burried deep within random bloggers who > didn't always update their tutorials to breaking changes, silverlight > forums and / or some random hack training you on "best practices for > Silverlight" which really had no official sponsorship from Microsoft. As a > Product Manager all I had was one question "How experienced are my > audience? with Silverlight ...level 100 - 300 breakdowns" and i kept > getting confused "starry" eyed looks like "Why does that matter?" .. i > needed to know how deep the features were being used, what issues tooling > is having with features, which features should we keep investing in and > which ones should we depricate etc.. but like most things at Microsoft it > was "Oooh look Shiney object.." (ie new release each 9 months). > > That being said, you have the capabilities to do a lot of plausible and > high performance driven solutions with but it always came back to "You > don't know what you don't know" and with Silverlight thats its weak legacy > but to say "it's brittle" or "it couldn't' do xyz.." you're going to have > to accept my pepsi challenge on that one as I see it much differently :) - > i was tempted to say "But you're doing it wrong" but i know how combative > that remark can get heheh :) > > > > --- > Regards, > Scott Barnes > http://www.riagenic.com > > > > On Thu, Feb 13, 2014 at 9:55 AM, Tony Wright <tonyw...@gmail.com> wrote: > > I was another who spend an awful lot of time learning Silverlight. Used it > on a major international project going out to a whole lot of clients where > we couldn't control the browsers that they use. We'd finally just got on > top of all the Silverlight quirks and MVVM when Silverlight was first > declared dead. So many hours invested. 1000s of hours per dev. We were > majorly peeved. > > > > But in hindsight, I understand. XAML looks great when done right, but it's > complicated and inefficient to code in. Real devs spent their time tweaking > the raw XAML. Not ideal. > > > > Then there's the brittleness of Silverlight. One little bug gets through > and the whole Silverlight app needs to be reset. There's no recovering by > hitting the back button either. It's nearly always a complete restart of > the app, and there's no guarantee you're not going to hit that point again. > > > > Debugging was crap as well. Oh, let's get fiddler out and see what's going > on because the app itself isn't showing anything in the exception handlers. > Duh. > > > > So now I've moved to MVC with Entity Framework and whatever flavour > Javascript library seems to be popular at the moment . Believe me, it's a > much happier environment to be coding in. You don't have to bitch as much. > One page might fail, but it doesn't bring the entire app down. > > > > As for the graphs - I'm using the Kendo (which is Telerik) graphs and data > visualisation tools. They're ok, and there are a couple of annoyances, like > with any graph generator, but they're pretty good. > > > > *From:* ozdotnet-boun...@ozdotnet.com [mailto: > ozdotnet-boun...@ozdotnet.com] *On Behalf Of *Scott Barnes > *Sent:* Thursday, 13 February 2014 10:31 AM > > > *To:* ozDotNet > *Subject:* Re: Migrating TFS > > > > Yes its 20years support (silverlight). Nearly all products get 20yr > support from Microsoft as it has something to do with overarching > Military/Govt contract agreements etc. > > > > I'm one of the people that's declared WPF/Silverlight dead and you will > not get an official response from Microsoft so you need to let go of that > idea aswell. > > Silverlight is dead as long as the plugin is installable and visual studio > can support its project(s) but dead as in future momentum / growth, yes. > It''s a Zombie, the body is still moving around but the brain isn't > functioning anymore. > > Just because you're working on Silverlight today doesn't mean anything, > I've got 10 guys right now working on WinForms but do we really want to > entertain the idea that WinForms is still relevant in future Microsoft > roadmaps or should we call it "dead" and move on. > > There is no alternative and that's why this crap we have HTML/JS is > getting beyond the magnitude of stupidity, as its like the ELSE statement > in the IF RIA == Alive logic, it's the retreat point to when good ideas go > bad and we have to say out loud "Well.. i guess we could go for breadth > user experience and ignore depth user experience" in our app development. > > Am I excited at the prospect that Silverlight has no future, no.. i > dedicated 3 years of my life to that product and i'm just as pissed if not > more pissed off about the stupidity of Sinofsky than probably most people > on the planet :) (in fact you can see my back and forth argument with Steve > on the weekend https://twitter.com/MossyBlog/status/432319248514289664) > > I suspect going forward if the rumours i'm hearing are true, that they'll > take the XAML runtime from Windows 8 and move the IP down to the Windows 7 > via an update or something to that affect. Basically they'll try and get > Windows 7 developers to start targeting the new UI namespaces in their UX > development which will unlock that bridge between Old and New...resulting > in getting Windows 8 pull through ... > > Now although that will suck initially as it won't help existing WPF/SL > solutions that use the old way of doing things it will however at least > start to unlock some more possibilities in that area. Having seen a years+ > development on WPF and Silverlight for some very expensive products here at > work (multimillion dollar deployments etc) I can't say it would be a > welcome solution but if they abandon the new namespaces for the existing > ones then the will also kill growth for Windows 8 - which isn't an option > especially with a new CEO. > > Again that's just spitball / speculation. > > > > > --- > Regards, > Scott Barnes > http://www.riagenic.com > > > > On Thu, Feb 13, 2014 at 8:20 AM, Greg Keogh <g...@mira.net> wrote: > > Hi Greg H > > > > I certainly agree that Silverlight is/was a great way to deliver > impressive apps in the desktop browser. Because it was XAML and C# I barely > had to learn anything new, I could sit down and churn it out (once you > knocked through all the security walls of course). I know you put a lot of > effort into Silverlight, we were all impressed with your timeline > visualisation. > > > > Does anyone know what the official lifetime of Silverlight is? Have > releases and updates simply stopped so that it will quietly go stale and > extinct on its own? Is there an official date for end of support? I ask > because we are still writing and releasing significant apps and customers > are going to ask us if Silverlight is dead (some already have apparently). > > > > What's the alternative to Silverlight for delivering browser apps with > rich graphics and charts? Have options improved in the last year? > > > > *Greg K* > > > > On 12 February 2014 17:54, Greg Harris <g...@harrisconsultinggroup.com> > wrote: > > I do not think this was directed at me but here goes... > > > > Start rant > > > > @#$%^ing Microsoft has #$%^&ed me and the community on Silverlight, I > spent a few years 100% focused on Silverlight at a significant cost in time > and money, all now just wasted! > > > > Today, I have a client that would 100% fit a Silverlight solution for > their line of business (LOB) application, but they are not willing to take > on Silverlight because of Microsoft's end of life perspective on the tool. > > > > I would agree that it may not be the right cross platform tool for all > mobile devices, but I see no reason why MS cannot make a commitment to > future releases and ongoing support on Windows, Mac, Windows Phone and > Android. > > > > I would not do the next version of Angry Birds with Silverlight, but I > would do most LOB apps with Silverlight. > > > > Microsoft, you have made me angry, you have made my client's angry, you > have lost credibility, I do not trust you! Probably more fool me for ever > trusting you! > > > > Microsoft, you could start to gain some credibility back by restoring > Silverlight to its rightful place as the tool of choice for client side > development in LOB apps with a commitment to maintain and support it for 20 > years into the future. > > > > End rant > > > > Regards > > Greg Harris > > > > On Wed, Feb 12, 2014 at 5:48 PM, Greg Keogh <g...@mira.net> wrote: > > Greg? Where are you? > > This is your cue. > > > > Ah! What! I'm awake ... I saw Silverlight mentioned as dead and > abandoned. Guess what I've been doing all day today .. expanding a large > Silverlight 5 app. We have no alternative, we've spent years developing the > app and it's in use by some gigantic companies internationally. > > > > What the hell else can we do? Seriously! Discussion here last year pointed > out that HTML5 is the only alternative to delivering rich apps on the > browser desktop, but it groans under stress and I was warned that it just > can't show attractive interactive charts of the type available with the > ComponentOne SL libraries. > > > > Also, I have subscribed to MSDN Magazine (MSJ as it was) since 1993 and I > agree that it is generally uninteresting these days because it's mostly > about JavaScript, Stores, Azure, Windows RT and Windows 8 (the latest > groovy stuff you're talking about). I find I flip through new issues and > chuck them aside. I like academic articles, but Petzold's and McCaffrey's > articles are so abstract they're in the twilight zone. > > > > My day to day development experience is consistently as infuriating and > unpredictable as ever. Projects won't build, IIS goes haywire with code > 500s, versions clash, dependencies are all over the shop, kits don't work, > samples are simplistic, designers crash, I'm coding XAML UIs by hand, I > have to learn WiX, I have to run VS2013 and VS2012 side by side due to COM > problems, my VS2013 is diseased, and so on. I get up in the morning and the > things that worked the night before are all on the fritz. Sometimes I miss > punch cards. > > > > However, I don't want to fuel the jovial atmosphere of impending doom that > pervades this forum ;-) > > > > Greg > > > > > > > > >
