| Visual Studio E... 的个人资料See Sharply日志列表 |  工具  帮助 |
|
4月6日 Another plugDan reporting,
I'm sure nobody even looks here anymore, but just in case...I've got a blog set up now: http://saiyr.blogspot.com/
I looked into Live Spaces, and unfortunately, I don't think I could deal with it.
Dan 12月12日 Shameless PlugWell it's December and I figured I should post how my life is going.
3rd Best Page in the Universe. It's my blog, check it out Maybe you'll like it and check back more often. 8月5日 The parting of the fellowshipSo Ed and Julia have already left. Michael is leaving tomorrow, and I on
Monday. So sad :'( Anyway, our code can be found here, some of which is still
unpublished: http://www.gotdotnet.com/codegallery/codegallery.aspx?id=e94bc8ef-87b1-424b-ae25-39e81004e86a
I've had an amazing summer, and I've met a lot of cool people (shameless
facebook plug http://purdue.facebook.com/p.php?id=13701676&l=47052cead2).
I'm definitely hoping to come back next summer. I don't have my own blog set up
yet, but I'm working on it, and posting that link will likely be my last post
here.
Dan 7月25日 Tuesday AfternoonSo the coding contest is over. It was a lot more fun than I expected it to be. Although I have not yet finished the puzzle I was working on, I can say that I was amazed at the complexity of it.
So I'm hoping that those of you that read this blog are technical people because the following is going to get very technical, and may even sound like rambling at some point. For this I apologize in advance.
So we started on Friday around 5pm. We get the spec off the website and start implimenting the virtual machine. So Luke works on his implementation and tests it. However, by some stroke of magic his program works the first time we run it. Needless to say this was amazing. At first we weren't sure that it was working though because we were getting this random garbage being printed out to the screen. But then we got smart. We decided to capture the output to a file. So we do that. And we open it in Visual Studio, and there's this message that says "Delete everything up to colon:" We do that and are left with this 15 meg binary file. Then we got curious and opened up that new file in the virtual machine. Lo and behold, we are prompted with a UMIX prompt. (Universal Machine IX.)
We log into the guest account and we're greeted with a message that says that some guest has been attempting to hack the system using this qbasic script. Yes, I did say qbasic. There was a nearly fully implemented version of Qbasic running in this UMIX OS. So we take the hint (afterall, it is a puzzle contest) and begin hacking away. However, there was an error? Oh no!. So we take a look at the hack.bas file and it turns out that the file was corrupted. We copy the file out into notepad and look at it to figure out what it does. It was a brute force hacking script. However, all the numbers were Roman Numerals! So we've already hit a crazy level of rediculousness and we're only like an hour into this thing. So after we obtain some passwords we start logging into those accounts while we continue to brute force the system and get as many passwords as we can.
I decided to login to the account named howie. I log in, and there is this adventure game. From what I was able to gather, adventure is a text based game that lets you examine, combine, destroy, and otherwise manipulate items while you walk around a map and complete some sort of task. I'm still working on the game now because as I discovered, the items are interdependent. So you need to figure out what item needs what thing and then combine them in the right order to make something. From what I can tell now, I am just over halfway finished with it. Some of the items have rediculous interdependices. Dan developed a tool to help analyze those dependencies. I also developed a tool that would find the items in a room and develop the query to examine all the items in the room. So once I had finished that tool, I began my attempt at solving the puzzle. Right now, I've got a script that runs through about half of the board and assembles everything for me. I just need to figure out the rest of the board.
There were also other puzzles. Michael and Karen were working on a bruteforcer that played this game for them and created the answers that they needed. Unfortunately, Michael only got it working right as the contest ended.
Luke and Anson worked on a ray-tracer in this crazy ASCII programming language called 2d. In this language there were ASCII boxes that represented functions and you had to tie them together to perform tasks. I really can't explain it much better than that, but it was crazy looking when they had it finished.
Cyrus and Tom worked on implementing a second virtual machine that was based off of this four operations that had intended operations, and unintended operations. It was called balance. The idea was to figure out what operations needed to be done in what order to perfom certain tasks and leave everything in a good state. It was crazy to listen in on some of the stuff they were talking about. It was a lot of stuff that I am only vaguely familiar with. (bitwise operations on registers in virtual machines.)
There was also the company picnic on Saturday where Michael, Karen, and I went and played volleyball in a tournament with some of her friends and a random girl that we picked up when we were there. we didn't advance anywhere in the tournament, but I played much better than I knew I could. I was actually surprised, because my serve was consisent. It was good to play well, and I got a nice tan. It is amazing to see how much darker I can get when I am only in the sunlight for a few hours. It was a good weekend.
So yeah, now I'm testing the latest add-in and making sure that it is stable. Our goal is to have everything done and ready to ship before Friday, so this is our final push to awesomness.
Hope you enjoyed the read. By the way, if you don't know the song, Tuesday Afternoon by Moody Blues. I like it, and so should you.
-ed
7月21日 Its Friday 3:48 PM, do you know where your coders are?So it is Friday, and the end of our 3 week push to get out some awesomeness. There has been all sorts of things coming out of our lab, the best part having our dif and merg come together in a couple days without too many problems. After struggling around for about a week and a half without a serious form of source control (e-mail and im before that) we finally got Team Foundation Server running and getting used to the idea of checking in and checking out code and making sure that everyone had the latest version. Once we got that working I started to merge everything together in the UI so that it was functional. However, my implimentation was a bit off so it was updated and now everything totally fits intop the look and feel of visual studio, except our use of a smiley face as an icon because we haven't picked anything new.
There is an interesting coding contest going on at http://icpfcontest.org that we are going to be participating in, along with some of our co-workers. I am looking forward to it because now I will be able to see how truly freakin awesome and smart the people that we work with are. You can tell when you talk to them, now we are going to get to see thim in action.
I still don't like my own brain because it takes forever to properly absorb the right things to do with the code. Once it all clicks I can do all sorts of things pretty quickly, but it takes forever for me to figure out what functions I need to call and that kind of thing. Thank God for some of the automatic completition things in Visual Studio. I would definitely be a lame coder without them.
Well, I guess in the next week, we can look forward to bug fixing and preparation of everything and making sure that everything is working well and all that stuff. Only two more weeks left before we all head back to our original places of living and continue on with our lives.
Oh, and it got really warm yesterday and today and it will continue with that. I wonder what that is about. Oh well, I guess the weather is still unpredictable. I bet the weather is just one of those things that the more we know the more we realize we don't.
Thanks for listening. That's all I've got for you.
-ed 7月18日 JourneyOk, since I didn't blog any last week, I'll try to make up for it with some solid information in this post.
Michael 7月15日 Nothing like working on a SaturdaySo our blog activity has taken a large dive over the past week. This is mostly due to the fact that we've started working on our big project, which we're supposed to be done with in a week. Me, being the optimist, think we've loaded up too much work to be finished so soon. The backend behind this add-in, which uses the code model to analyze differences between version of a file, is mostly complete. However, the UI, last I checked isn't, which was a M1 goal. I wasn't sure we'd get the actual merging operations done, and sure enough, we didn't.
Since I assigned these tasks to two others to work on, I was worried about writing unit tests for the code I'd already written. This turns out to be a problem, since the entirety of CodeModel is a bunch of interfaces which are COMObjects when you actually use them in Visual Studio. Thus, I had to come up with a way to write automated unit tests. I ended up implementing all of the interfaces that I needed, but only wrote the methods that I needed--a dire, boring, tedious task. So dire, so boring, and so tedious that I pretty much decided that I didn't feel like working Friday afternoon, around 2. Needless to say, that's why I'm in the office right now.
Today I got some unit tests working (12% code coverage! Hurray!), and might work on it some more later on, depending on whether or not I feel like it.
Dan 7月9日 Bring your pardner to work day!Once upon a time Julia decided to go to work on a Sunday! Why?! Because she has a crummy laptop and she wanted to play the sims but she didn't have the install cd's (she must have left them at home or in her dorm) so she was really bummed but then she decided to go to work and drag her pardner along with her!
For no apparent reason he decided to show up! They walked around looking for hard workers that Microsoft who deserve to be paid more because they love work so much! There was this random guy on the second floor, an Indian guy we didn't know, Eric Maino and Roc Yu! They like to come in on Sundays and do extra work when no one's looking!
Come visit us and we'll share HERSHEY'S PEANUT BUTTER KISSES!!!
-Team McAwesome (Julia and Dan)
 Arrrrgh! Ahoy mateys!Yeah, so we all went to see Pirates of the Caribbean: Dead Man's Chest on Thursday night/Friday morning.
-Julia 7月7日 Wrap up, Start upMy parents came up last week and it was cool to do a lot of the more tourist type activities in Seattle. We also drove across the state to Spokane to meet up with some relatives from Montana who we don't get to see very often. It was crazy to see how the landscape went from green, snow-capped mountains, to Grand Canyon-esque dry cliffs, to Kansas plains as you drive east. We also took the ferry to Victoria, Canada which was a cool place.
On the work side of things, we're trying to get our add-ins wrapped up and released. Which makes me wonder why I started to try and add more features into the Jump To one today. I didn't finish and to really polish up the new feature it'd take a bit more work and I'm weary of that because we are diving in to our main Merge Documents type project on Monday. I did fix the bug Anson found when he was trying it out yesterday, but I might just bail out and delete the stuff from what I was working on today (trying to get it to dynamically update the list as the user adds and deletes code). So they might end up just having to close and reopen the window. The main problem I can't figure out how to deal with is the ElementChanged event. If they rename an item, how can I look through my list of item names and know which one to change since it doesn't pass in the old state of the item? I thought I had the answer with the ElementID property, but it's not implemented in C#. The other issue that kind of frustrated me is that I was trying to speed the add in up some by switching some things over to SortedLists<> and changing the way some things were implemented, but I didn't realize until about an hour after I started working on it that the keys needed to be unique. I was just using the Element names. I guess I was hoping it'd let me specify a secondary comparer if the elements had the same original key, and return an array of items when I searched for a key, but it didn't. Should have noticed that sooner, oh well, probably still a bit tired from the midnight showing of Pirates of the Caribbean last night. I think I'm just gonna try to wrap it up and call it done except for bug fixes people find next week. However the main focus will be starting on our new project.
-Michael We have roles!....sort of. We are starting our final project for our internship. It is ambitious, but I am confident that we can get it done. For now Michael and Julia are going to be doing more PMish things, Dan is the Dev lead, and I am the Test lead. These roles are probably a bit arbitrary and subject to change as we'll all be doing a little bit of everything.
!!!SPOILER ALERT!!!
We went and saw Pirates of the Carribbean last night. I thought it was good, though I am dissapointed that I have to wait until next summer to see how the story resolves.
!!!END SPOILER ALERT!!!
Yeah, we're still working on getting our add-ins to release them to people. I believe we've started dogfooding them and test them on other people's machines. Hopefully if there's any oustanding issues they can be resolved, but right now I think everything is in a good state.
-ed 7月3日 July 03 1:37 PMSo on Friday we gave a presentation showing off the latest wares that we have created in our See Sharply Studios. The PMs we showed it too looked excited about what we had done and we're looking to release them sometime soon (those details are being worked out.)
So I came in today to work on my project proposal some. It's tough because I keep thinking that every idea I come up with is hopelessly lame. Hopefully I can have something working that is actually kinda cool that doesn't turn up being totally dumb. Then again we may end up working on something entirely different, depending on the quality of the proposals coming from my fellow explorers.
And in other news, taxes suck. I think I am going to propose an ammendment to the constitution repealing the income tax and replacing it with a national sales tax or something. (Too bad the government can't work for free.)
Yeah, that's all I have right now.
-ed 6月28日 Thoughts from a bored internWell, it's 10:30 at night and I'm pretty bored, and since I fell off the front page again (no thanks to my "pardner"!), I figured a blog post would be a good idea. I believe all three addins have been completed, and we may or may not be presenting them tomorrow to the VCS PMs. Let's just hope that the programs are stable enough and don't BSoD during the presentation! I spent a number of hours trying to figure out the extensibility APIs, and discovered that using the text editor methods were more complex than one would think. The addins that Team McAwesome wrote dealt a lot with the TextSelection class, which proved to be an interesting learning experience:
Dan 4weeksToday signifies four weeks into our internship. And for it we have 3 navigation add-ins for Visual Studio 2005 and an internal website that we've made. It's been a fun time.
Yesterday we went to see Superman as a morale event. It's a good movie. If you haven't seen it yet, you should.
What I've realized in the last four weeks is that C# and Java are much easier to develop Object Oriented programs than C++. I guess that's because C# and Java were designed for use in Object Oriented environments much more than C++ was. I also like that in C# and Java there is less worry with memory management as that is taken care of by their respective virtual machines. I am beginning to like C# more than Java because C#'s syntax appears to be more intuitive and has less of a learning curve than Java.
We had a mild heat wave here over the weekend where it got to (gasp!) 92. It didn't feel as hot to me because it wasn't accompanied by the high humidity that I am used to back home on Long Island, but the people on the local news said it was record highs. This place is pretty cool in the summer. since it was warm, some of the interns went out to a local park and played some volleyball and ultimate frisbee. It was nice to get outside and enjoy the warm weather.
I set up a website for myself, www.edflyer.com. There's nothing there right now, but I'll eventually have stuff up there even after I have left Microsoft, and probably a mirror of this blog.
A review of BLINQ that I found. From Tim Heurer. Has a follow-up.
-ed
6月27日 Why comment? My code reads like poetrySo except for a tiny bit of tidying up, our add ins are all finished I believe. Honestly, I think they are pretty kickin' so I might have to launch a personal ad campaign to spread the word and get people using them.
Looking back on a project, I'm often surprised at which aspects turned out to be the most difficult and which were the easiest. For example, for our navigation add in, I definitely thought generating a list of code elements to jump to would be the most difficult part. However, VS already had functionality built in to identify code elements and their type and so not only was it just a simple search to generate a list, it was easy to filter it based on any criteria you want. A little more work and you can add some icons to the list as well and have your own tool looking as slick as IntelliSense.
I definitely did not think the most difficult part would be integrating the tool as a tool window into VS, but it was, by far. It's not so much difficult to do as it is difficult to learn how because it's not a well documented area. So I'm hoping people will actually dock the window occasionally to actually make it worth the trouble.
There were a few other behind the scenes changes that had to be made when switching it to a tool window that I thought were interesting. When it was still a standalone windows form, basically you would execute the command associated with the addin and it would pop up, generate its list, etc... and then you'd close it and it would be done executing until the next time you ran it by pressing the keyboard shortcut or clicking it in the menu. However, you generally want your tool window to load on startup. So the add in had to change so that when the command is executed it basically toggles its visibility, but we had to add the ability for it to update its list since most often a solution wouldn't be open when the add in is first loaded.
Guess I'll write some comments now.
Oh, and I've acquired speakers. :) 6月23日 DLinqHOLI broke it!
I was supposed to have the DLinq Hands-On Lab done today, and I was really making progress on it this week. I was supposed to finish the last part, "Working with Advanced features", today but then SqlMetal decided not to work. I don't know what I did to break it but I definitely don't understand what's going on. "SqlMetal has encountered a problem and needs to close. We are sorry for the inconvenience."
Aside from that, the Hands-On Lab was pretty helpful. It took me a lot more than 90 minutes to go through it because I don't know C# and I wanted to know what it was doing. I guess I have more of a grasp on what linq actually does and it'd probably mean more if I knew anything about SQL. It really annoys me when people say "Don't worry, SQL's easy, you'll just pick it right up" because it won't make any sense to me until I have to use it.
One of the greatest resources I've discovered is //mylearning. Ok, so I didn't really discover it, one of the PM interns from MIT was going through it and he wanted a questions buddy. I finished the "Functional Specifications at Microsoft" online course and learned all about writing specs. I thought PMs only went to meetings but writing specs is totally right up my alley. Whenever I do projects (yes, even on my own) I always outline everything (objectives, steps, possible issues, which parts are being graded) and I make little notes. I can't believe you guys get paid to do that.
In other news, I think I just got roped into playing volleyball at the 2006 company picnic. And watch out for team snakes on a plane for puzzleday.
-Julia 6月22日 Team McAwesomeGroup 1 (Daniel and Julia a.k.a. TEAM MCAWESOME):
Move to next/previous method
-Through two simple shortcut keys, this feature allows a user to quickly navigate to the next or previous method given their current cursor position. Collapse comments -This feature collapses all comment regions in the current editor Another exciting morning of classes later, we went on the MS Home tour. Remember the movie Antitrust? With Ryan Phillippe? It was a lot like the house from the movie. I like the RFID tags idea but it all seems a little too creepy ... I don't like the idea that my house can buy stuff for me according to my tastes but can't clean itself (now that would be useful). Also, if your kids grow up with so much automation, they're going to be completely completely dependent on it. Right now I'm so dependent on MS Word that I don't think I can write a paper by hand (spellcheck, drag and drop, thesaurus--I've never actually used a printed thesaurus and I'm not sure I'll know what to do with it if I had one, THE BACKSPACE BUTTON, find all/replace all, and all the spacing modifications I have to make so that my papers are long enough). I'm not saying handwriting 20-page papers is necessarily a useful skill but simple tasks like grocery shopping and making dinner are pretty useful.
I really like the idea of being able to adjust music, lighting and temperature from the couch though. Right now the remote control for my tv at home is missing, so I kind of have a voice-activated device in place. It's not very efficient though because she has to stand up on her tippy toes to reach the buttons and she doesn't always do what you ask her to. Little sisters can be pretty useful if you train them right.
Team McAwesome went to Wendy's today. Team Not-So-Awesome didn't want to. Laaaame.
-Julia 6月21日 My problems with the 'Find' tool in browsersSo I gotta admit, writing add-ins to Visual Studio is much easier than I thought it'd be originally. I think it's really just another example of what Luke and I discussed some when I was shadowing him yesterday: that often the hardest part of diving into a new API/SDK/whatever is just getting something to show up on the screen. After that it's pretty smooth sailing. The documentation on the MSDN was more useful than I thought it'd be too. The tutorials are a little skimpy, but they atleast give you a sense of what classes you'll be using and need to read up on some more.
So first, a brief description of what the add-in actually does. Basically it's a navigation feature that lets the user pull up a small window with a keyboard shortcut. The window displays a list of classes, functions, variables, etc... that are in your current solution, sort of similar to IntelliSense. As the user begins to type the list gets filtered and sorted... actually very similar to the search box on facebook.com for tagging a person in a photo if you happen to be familiar with that. Then you can press enter and the IDE will navigate to where that class or function is implemented so you can edit it. Basically it's like being able to easily call 'Go to Definition' on something without actually having to put your cursor over a reference to the object.
Now like I was saying, progress went really well, so Ed had an idea to expand it so that the user could also tab through references to the object instead of just going to where it's defined. Initially I thought that would be pretty cool, but after some thought I've decided I don't think it'd be a good addition to this add-in. My main concern is with disorienting the user. One of the main ideas behind our navigation tool is that its very unintrusive and won't take up much screen space, you just pull it up, type a few letters, press enter, and you are taken to a new place in code. Now with my initial impression this idea wouldnt' change that. However, if I step through a scenario of myself using it... Let's say I pull it up and type 'myf' which is enough for it to filter to 'MyFunction'. I press enter and am taken there. The first obvious change is that with this addition the user would have to take an extra step to close the window because it wouldn't know if the user wanted to go here and stop or tab through to see where it is referenced. Maybe you could have any keystroke other than 'tab' dismiss the window after an initial selection is made, but that's not my main concern. Let's say that the user decides to tab through and see the different references. This would basically be like how 'Find' is implemented in many word processors and internet browsers. You type in a word and hit 'Find Next' and it just keeps taking you to the next one. The main problem I have with systems like that is that there is no way to know how many instances of that word there are in the document unless you count them. And often I don't even realize I've looped back to the top of the document until two or three clicks later. And let's say it jumps me to page 33 of some PDF file. I have to spend a few seconds looking around at the scroll bar, page number, and chapter headings just to get my bearings for where it's taken me. I imagine the same thing would happen to users tabbing through references to an object. With no indication of where the references are or how many of them exist, the user would have to spend a few seconds at each jump to orient themselves becuase they'd really have no idea where it'd take them next. We could add stuff onto the visual display for the tool I guess, but I'd prefer to keep it looking more like IntelliSense with a text field stuck on top instead of a full blown window. Furthermore we would end up basically duplicating the functionality of 'Find All References' and it has a nice window that lists each reference with their location so the user can simply double click or F8 their way through them to find the appropriate one. So I'll go with the keep it simple philosophy on this one.
-Michael 6月20日 Tuesday's InformationWe started out our second round of projects today. We have been split into two groups that are doing add-ins for Visual Studio. We'll probably put them up on msnd once they are all ready and done.
I shadowed another Dev yesterday. It was cool. I found out that the code base for Visual Studio is very large and that it takes a good while before you can be familiar with it before you can start working features for Integration. I think I like the developer and PM ends of the software sprectrum much more than QA/Test. I like creating things and building them, which is why I would like being a DEV. I would like to do PM work because I like discussing ideas and figuring out what can be done and what is too costly. Those are the impressions I've gotten so far from my shadowing expereinces.
Our PM finally has her blog up, and since she posted a shout-out to us, I feel that it is only appropriate that we post a shout-out back. You can check out Karen's blog, Karen's Space here. She just got back from Tech Ed '06, and from what I understand, she did very well. She "owns" the C# IDE experience and her blog reflects that. She's got a list of keyboard shortcuts that can help you be more productive in Visual Studio while developing in C#. Check it out and send her feedback.
-ed 6月19日 Batter up?!Well, I suppose I'm up to bat again. The past week has mostly been shadowing for all of us. So far, I've shadowed a PM, two SDEs, and a SDET. I'm now further convinced that being a PM would not be in my best interests (because I'd likely be terrible at it), and I still find it rather interesting that I'm the only person on this team that doesn't want to be a PM. As far as the other two positions go, I think I'm about 70/30 for SDE/SDET, rather than flat out 50/50 like I was before last week. I haven't followed up on the Haskell link I posted at all, since it's blown my mind a number of times now, but maybe I will some day (read: won't ever). Class is the same old boring C++ stuff, speaking for myself, so I haven't really paid much attention, though I know we covered virtual functions today at the end of class. Our next project(s) will be decided today. I think we'll end up working in pairs, which may be interesting. Dan |
|
|
