Search Results

Search found 28672 results on 1147 pages for 'best practise'.

Page 342/1147 | < Previous Page | 338 339 340 341 342 343 344 345 346 347 348 349  | Next Page >

  • Central renderer for a given scene

    - by Loggie
    When creating a central rendering system for all game objects in a given scene I am trying to work out the best way to go about passing the scene to the render system to be rendered. If I have a scene managed by an arbitrary structure, i.e., an octree, bsp trees, quad-tree, kd tree, etc. What is the best way to pass this to the render system? The obvious problem is that if simply given the root node of the structure, the render system would require an intrinsic knowledge of the structure in order to traverse the structure. My solution to this is to clip all objects outside the frustum in the scene manager and then create a list of the objects which are left and pass this simple list to the render system, be it an array, a vector, a linked list, etc. (This would be a structure required by the render system as a means to know which objects should be rendered). The list would of course attempt to minimise OpenGL state changes by grouping objects that require the same rendering operations to be performed on them. I have been thinking a lot about this and started searching various terms on here and followed any additional information/links but I have not really found a definitive answer. The case may be that there is no definitive answer but I would appreciate some advice and tips. My question is, is this a reasonable solution to the problem? Are there any improvements that I could make? Are there any caveats I should know about? Side question: Am I right in assuming that octrees, bsp trees, etc are all forms of BVH?

    Read the article

  • Need to hookup HP dv7-3085dx with Nvidia GeForce GT 230M to my Dell 30 inch LCD 3007WFP at max resol

    - by user14660
    I recently bought an HP laptop (dv7-3085dx) (http://reviews.cnet.com/laptops/hp-pavilion-dv7-3085dx/4505-3121_7-33776108.html) which is supposed to have a pretty good video card (NVIDIA GeForce GT 230M). The card is supposed to output a max resolution of 2560x1600 which is also the max resolution of my monitor (http://www.ubergizmo.com/15/archives/2006/02/dell_3007wfp_on_dell_2001fp_action_8_megapixel_desktop.html). Now I bought an HDMI to dual link dvi (http://www.amazon.com/gp/product/B002KKLYDK/ref=oss_product) cable...this is after Best Buy's 70 dollar hdmi to dvi (perhaps it was 'single' link?) didn't give me the best resolution. In windows 7, when I try to set the max resolution for my 30 in monitor, I only get 1280x800...which is absurd. The monitor is great, I love the laptop and the video card supposedly supports such resolutions. So I can't figure out why I'm not getting better resolution (by the way, when i "detect" my monitor in windows 7, it is shown correctly as DELL 3007WFP!).

    Read the article

  • Security tips for adding wireless AP to domain network?

    - by Cy
    I am researching best-practices for adding wireless to our existing domain network. My DHCP server is running Windows Server 03 Standard (not sure if thats useful). I am familiar with simple home networking but I thought I'd get some expert advice for the more advanced stuff. Any tips and / or best-practices? Is this Cisco Wireless Access Point a good option? Are there any additional hardware recommendations? Thank you in advance for your help.

    Read the article

  • Is there a constraint-based scheduling/calendar application?

    - by wonsungi
    Is there a constraint-based scheduling/calendar application? This application would be used to coordinate multiple people's schedules. Two basic use cases: Multiple people need to schedule a time to meet together. Everyone is busy at different days/times. Each person enters blocks of days/times they cannot meet, and the application suggests the best times to meet given a desired time range. Multiple people need to use some common resources for a specific length of time (over some time span like a week), but the exact date/time does not matter. These people enter the resources and time needed, and the application suggests the best way to share these resources. This use case still accounts for people's blocks of busy time. I imagine this program would be graphical, but other interfaces would be acceptable. Also preferable if web-based/works on both PC's and Mac's, but PC-only/Mac-only solutions are acceptable.

    Read the article

  • Which file format to use for simple video editoring

    - by Craig HB
    I've just bought a Sanyo VPC-CG10 camcorder which saves clips as MPEG-4 (specifically, High-Definition 720p MPEG-4 AVC/H.246) I want a simple, cheap way to edit the clips and as, I'm using Windows Vista, I though Windows Movie Maker would be a good choice. But you can't import mp4 files into Windows Movie Maker. I've tried various codecs, but I'm not sure what is the best format to convert to, so that I can edit the files in Windows Movie Maker. AVI, MPG, WMI... I'm not sure what format is best. Should I try to keep the movie clips in MPEG-4 and edit them in that format (and then which editor do I use?) or is there a better format that I should convert to before editing (and what format is that?) ?

    Read the article

  • We are moving an Access based corporate front-end into a Web-based App

    - by Max Vernon
    We have an enterprise application with a front end written in Microsoft Access 2003 that has evolved over the past 6 years. The back end data, and a fair amount of back-end logic is contained within several Microsoft SQL Server databases. This front end app consists of around 180 forms, and over 120,000 lines of code, and interacts with VB.Net DLLs that support various critical functions used by our sales force. The current system makes use of 3 monitors to display various information; the Access app uses COM+ to control Microsoft Outlook and Internet Explorer for various purposes. The Access front end sometimes occupies 2 screens, automatically resizing itself based on Windows API-reported screen dimensions. The app also uses a Google map to present data to our agents, and allows two-way interactivity with the map through COM+ connectivity to JavaScript contained in the Google map. At the urging of senior management, we are looking to completely rewrite this application using some web-based technology, such as ASP.Net or perhaps a LAMP stack (the thinking with the LAMP stack thing is "free" is pretty cheap). We want to move to a web-based app so we can eliminate the dependency on our physical location for hiring new sales force members. Currently, our main office is full to capacity, and we need to continue growing the company. Does anyone have any thoughts on what would be the best technology to use for a web-based app of this magnitude? Keeping in mind the app is dependent on back-end services on our existing infrastructure. The app handles financial data and personal customer data, among other things. [I've looked at Best practices for moving large MS Access application towards .Net? and read the answers, and most of the comments. Interesting reading, and has some valid points, but our C.O.O. and contracted Software Architect are pushing for a full web-based app, not a .Net Windows App]

    Read the article

  • Domain Transfer Protection - need advice

    - by Jack
    Hey, I am about to purchase a domain name for a bit of money. I do not personally know the person who I am purchasing the domain name from, we have only chatted via email. The proposed process for the transfer is: The owner of the domain lowest the domain name security and emails me the domain password, I request the transfer After the request, I transfer the money via PayPal When the money has been cleared the current domain name owner confirms the transfer via the link that he receives in that email I wait for it to be transferred. The domain is currently registered with DirectNIC - http://www.directnic.com/ Is this the best practice? Seeing I am paying a bit of money for this domain name, I am worried that after the money has been cleared that I won't see the domain name or hear from the current domain name owner again. Is there a 'domain governing body' which I can report to if this is the case? Is the proposed transfer process the best solution? Any advice would be awesome. Thanks! Jack

    Read the article

  • How to copy only VM changes in VMware to another physical machine

    - by Paul
    How can I use VMware Workstation to copy only changes to a VM to another physical machine, rather than recopying the entire VM every time? Can I just copy the snapshot files, if the second physical machine starts from a pristine copy of the VM? Would it be best to create a linked clone against the original VM, and then copy the linked clone directory each time? Does that even work? (I'm assuming I'll have to change the path in the VM's metadata file, since the paths will be different. I'm assuming I can't clone a linked clone w/o cloning the base.) Background: I and another developer with VMware Workstation are using VM's for systems development. I have to work from a home desktop. Transporting a multi-gigabyte VM either direction takes several hours at best. Downloading it is easiest, but that takes an hour to get it on a DMZ machine for download and then many hours to actually download.

    Read the article

  • Help parsing long (3.5mil lines) text file, line by line and storing data, need a strategy

    - by Jarrod
    This is a question about solving a particular problem I am struggling with, I am parsing a long list of text data, line by line for a business app in PHP (cron script on the CLI). The file follows the format: HD: Some text here {text here too} DC: A description here DC: the description continues here DC: and it ends here. DT: 2012-08-01 HD: Next header here {supplemental text} ... this repeats over and over for a few hundred megs I have to read each line, parse out the HD: line and grab the text on this line. I then compare this text against data stored in a database. When a match is found, I want to then record the following DC: lines that succeed the matched HD:. Pseudo code: while ( the_file_pointer_isnt_end_of_file) { line = getCurrentLineFromFile title = parseTitleFrom(line) matched = searchForMatchInDB(line) if ( matched ) { recordTheDCLines // <- Best way to do this? } } My problem is that because I am reading line by line, what is the best way to trigger the script to start saving DC lines, and then when they are finished save them to the database? I have a vague idea, but have yet to properly implement it. I would love to hear the communities ideas\suggestions! Thank you.

    Read the article

  • PostgreSQL RAID configuration

    - by Yoldar-Zi
    I'm stuck how best to configure disk array. We have Hp P2000 G3 disk array with 24 SAS physical disks 300Gb each. We need to configure this array got 2 copies of PostgreSQL 9.2 because two different system. As we know it's recommended to store database and transaction logs(pg_xlog) files on separate disks. So we must setup 4 logical disk: 2 for transaction logs with RAID 1 2 for database with RAID 10 Is this right scheme of distribution? Or may be it is best to just make one big RAID 10 with 4 logical disks?

    Read the article

  • Effective handling of variables in non-object oriented programming

    - by srnka
    What is the best method to use and share variables between functions in non object-oriented program languages? Let's say that I use 10 parameters from DB, ID and 9 other values linked to it. I need to work with all 10 parameters in many functions. I can do it next ways: 1. call functions only with using ID and in every function get the other parameters from DB. Advantage: local variables are clear visible, there is only one input parameter to function Disadvantage: it's slow and there are the same rows for getting parameters in every function, which makes function longer and not so clear 2. call functions with all 10 parameters Advantage: working with local variables, clear function code Disadvantage: many input parameters, what is not nice 3. getting parameters as global variables once and using them everywhere Advantage - clearer code, shorter functions, faster processing Disadvantage - global variables - loosing control of them, possibility of unwanted overwriting (Especially when some functions should change their values) Maybe there is some another way how to implement this and make program cleaner and more effective. Can you say which way is the best for solving this issue?

    Read the article

  • relationship between the model and the renderer

    - by acrilige
    I tried to build a simple graphics engine, and faced with this problems: i have a list of models that i need to draw, and object (renderer) that implements IRenderer interface with method DrawObject(Object* obj). Implementation of renderer depends on using graphics library (opengl/directx). 1st question: model should not know nothing about renderer implementation, but in this case where can i hold (cache) information that depends on renderer implementation? For example, if model have this definition: class Model { public: Model(); Vertex* GetVertices() const; private: Vertex* m_vertices; }; what is the best way to cache, for example, vertex buffer of this model for dx11? Hold it in renderer object? 2nd question: what is the best way for model to say renderer HOW it must be rendered (for example with texture, bump mapping, or may be just in one color). I thought it can be done with flags, like this: model-SetRenderOptions(RENDER_TEXTURE | RENDER_BUMPMAPPING | RENDER_LIGHTING); and in Renderer::DrawModel method check for each flag. But looks like it will become uncomfortable with the options count growth...

    Read the article

  • How do you persist installed software & configurations on an Amazon EC2 instance?

    - by Richard
    I've gotten a base Debian AMI up and running and now I need to know the best way to maintain it. I've ran the updates (aptitude update/upgrade) and installed/configured my software (Apache, Ruby, etc.) but if I reboot the instance or start a new one I'll have to do all this work over again. How do you persist these types of things over a reboot? Do you build a new AMI every time you adjust some tiny piece of the system? Or is there some way to feed it a script on startup that configures it in "real-time"? I know I could go all the way with a Reductive Labs Puppet style setup but that's a bit too much for my needs right now (1-2 servers). Any best practices on this? Update: I found a bit of information on using User-Data to run scripts at instance boot time.

    Read the article

  • lamp server permissions on development server

    - by user101289
    I run a LAMP server on a ubuntu laptop I use only for development. I am not greatly concerned with security, since the server is never accessible outside the local network, and it's turned off when I'm not using it. My question is what is the simplest and 'best' way to set permissions/users/groups so that when my myself user creates, edits or writes files in the webroot, I won't need to go through and CHMOD / CHOWN everything back to the www-data user? Should I add myself to the www-data group? Or chown the webroot to www-data:myself? Or is there a best practice for this situation so I don't have to keep re-setting the ownership of these files? Thanks

    Read the article

  • What hash should be used to ensure file integrity?

    - by Corey Ogburn
    It's no secret that large files offered up for download often are coupled with their MD5 or SHA-1 hash so that after you download you can verify the file's integrity. Are these still the best algorithms to use for this? Obviously these are very popular hashes that potential downloaders would have easy access to. Ignoring that factor, what hashes have the best properties for being used for this? For example, bcrypt would be horrible for this. It's designed to be slow. That would suck to use on your 7.4 GB dual layer OS ISO you just downloaded when a 12 letter password might take up to a second with the right parameters.

    Read the article

  • Windows server detected error with hard disk

    - by user53864
    We have hosting Windows server 2008 R2 and I am working as admin in small company. The server is hanging and restarting as the hard disk seems to be damaged due to power fluctutaion(though having inverter) as it's showing the below error message on server reboot: Problem detected with the hard disk Press any key to continue It's Seagate 1TB SATA hard disk and it's booting after pressing enter. So it's clear that the hard disk is dying. Yes, it's in warranty but it's fact that warranty won't recover the lincesed windows server 2008 and it's data. As it's booting now, I backed up required things and I am thinking to clone the entire hard disk. The first thing it striked me is checking on the Seagate site if any tool available for cloning and I found Seagate DiskWizard but not specified it for windows server 2008. Please anybody could help me giving your best ideas for the below: Urgently, What's the best way(free of cost) for me to clone in my case with the new same sized hard disk? It's a one time lincenced and I cannot use the same key again if I reinstall the server. Will the lincense be carried with new disk if cloned? else there is a way to contact Microsoft explaining the problem occurred, to obtain new key for no charge?. I want to take measure for future. How do I keep two disks in continuous sync? mirrored & raid are the only options converting the disks to dynamic? or is there a best way I could do with no additional charge?. Any help is greatly appreciated. Thank you! EDIT:1 I started cloning the disk with CloneZilla and it was going proper showing in GUI. But after some time there is no GUI but a black screen with some codes(looks like disk location numbers) going page by page(I have attached the screenshots below captured from my phone). Do you people think it's actually cloning?. I started in the morning and it's evening now. I left the office now to let it finish what it's trying to do and I'll go & check it tomorrow. Slowly lost hope, don't know what face it's going to show tomorrow. Any ideas?

    Read the article

  • Multilingual sites and Google search results, using sub-folders for language

    - by AWinter
    About three months ago we added an English version of our, previously Japanese only, site under the subfolder /en/ we've tried to follow the sometimes incomplete best practices laid out by Google by adding alternate tags to all pages that are currently translated. The top page for instance has the following meta tags for language. <link rel="canonical" href="/"> <link rel="alternate" hreflang="ja" href="/"> <link rel="alternate" hreflang="en" href="/en/"> While the English main page under /en/ has <link rel="canonical" href="/en/"> <link rel="alternate" hreflang="ja" href="/"> <link rel="alternate" hreflang="en" href="/en/"> Alternate languages are setup in the sitemap. (as per Google's recommendations) It seems however that Google absolutely refuses to show the English top page in results when the user is using English at google.com if you search you'll, as of this post, get the Japanese description and a title that Google has apparently invented instead of the title and description in the meta-tags for the /en/ index page. Does anyone have any experience with subfolders actually working to affect search results? What are the best practices for ensuring that the correct language version of my website is displayed through Google and other search engines? And how long will it take before the new language version becomes prominent in search engine results?

    Read the article

  • Experienced programmer, beginner at web design, tools for effective maintainable web design? [closed]

    - by Clinton
    I do quite a bit of programming in my work, which I'm comfortable with, but recently I've being trying to do some web-design for non-work related reasons. I've got a Drupal site up and running, and added some content. But they all look fairly basic. Header with some content. It doesn't look particularly polished. Anyway, as an example, what I wanted to do was make some "bubbles", each with some text in them. From a programmers point of view, say: bubble(question_text, answer_text) might expand to a box with some border, with "Question: " + question_text then "Answer: " + answer_text. Of course I'd have lots of these bubbles, but I'd like to change their look and feel in one place, so simple HTML would be a maintainable nightmare. I also want to lay them out on the screen in some fashion. I was thinking a mixture of javascript and CSS, or possibly use PHP which Drupal uses. On the other hand, I fear I might be taking a 1990s approach to this, and that there's actually tools available now that make this process a lot easier. I'm just wondering what the best approach to this sort of task is? Should I be using offline web design software and copying the code to Drupal, and if so, any recommendations? I'm sorry if my question is a bit vague, because I'm not really sure what question I should be asking. I'd appreciate if you answer and comment, and I'll try my best to be more specific as I understand more.

    Read the article

  • Problem with a* implementation in pygame

    - by piyush3dxyz
    Yesterday i decide to make RTS game in pygame(pygame is best).I figured out many components of RTS game like unit selecting,health,resources but only 1 thing i still not understand.. which is a* pathfinding in pygame... I also done little bit of research on wiki,articles and papers...but still cant figure out problem.... function A*(start,goal) closedset := the empty set // The set of nodes already evaluated. openset := {start} // The set of tentative nodes to be evaluated, initially containing the start node came_from := the empty map // The map of navigated nodes. g_score[start] := 0 // Cost from start along best known path. // Estimated total cost from start to goal through y. f_score[start] := g_score[start] + heuristic_cost_estimate(start, goal) while openset is not empty current := the node in openset having the lowest f_score[] value if current = goal return reconstruct_path(came_from, goal) remove current from openset add current to closedset for each neighbor in neighbor_nodes(current) if neighbor in closedset continue tentative_g_score := g_score[current] + dist_between(current,neighbor) if neighbor not in openset or tentative_g_score <= g_score[neighbor] came_from[neighbor] := current g_score[neighbor] := tentative_g_score f_score[neighbor] := g_score[neighbor] + heuristic_cost_estimate(neighbor, goal) if neighbor not in openset add neighbor to openset return failure here is the pseudocode for wiki a* implementation......

    Read the article

  • How to make the jump from consumer support to enterprise support?

    - by Zac Cramer
    I am currently a high level consumer break/fix technician responsible for about 300-400 repairs a month. I am good at my job, but bored, and I want to move into the enterprise side of my company, dealing with Server 2008 R2 and exchange and switches and routers that cost more than I make in a month. How do I make this transition? Whats the best thing to learn first? Is there a standard trajectory for making this leap from consumer to business? I am full time employed, so going back to school is not a great option, but I have no life, so spending my nights and weekends reading and practicing is totally within my realm. I am basically overwhelmed by the number of things to learn, and looking for any advice you may have on the best way to proceed. PS - I apologize if this is a not quite the right forum for this, I know its not a technical question exactly, but I also know the sorts of people I want to answer this question are reading this website.

    Read the article

  • Hash Algorithm Randomness Visualization

    - by clstroud
    I'm curious if anyone here has any idea how the images were generated as shown in this response: Which hashing algorithm is best for uniqueness and speed? Ian posted a very well-received response but I can't seem to understand how he went about making the images. I hate to make a new question dedicated to this, but I can't find any means to ask him more directly. On the other hand, perhaps someone has an alternative perspective. The best I can personally come up with would be to have it almost like a bar graph, which would illustrate how evenly the buckets of the hash table are being generated. I have a working Cocoa program that does this, but it can't generate anything like what he showed there. So the question is two fold I suppose: A) How does one truly interpret the data he shows? Is it more than "less whitespace = better"? B) How does one generate such an image based on some set of inputs, a hash, and an index? Perhaps I'm misunderstanding entirely, but I really would like to know more about this particular visualization technique. Or maybe I'm mis-applying this to hash tables rather than just hashes in general, but in that case I don't know how it would be "bounded" for the image.

    Read the article

  • Spreading incoming batched data into a real-time stream

    - by pr1001
    I would like to display some events in 'real-time'. However, I must fetch the data from another source. I can request the last X minutes, though the source is updated approximately every 5 minutes. This means that there will be a delay between the most recent data retrieved and the point in time that I make the request. Second, because I will be receiving a batch of data, I don't want to just fire out all the events down a socket once my fetcher has retrieved it: I would like to spread out the events so that they are both accurately spaced amongst each other and in sync with their original occurrences (e.g. an event is always displayed 6 minutes after it actually happened). My thought is to fetch the data every 5 minutes from the source, knowing that I won't get the very latest data. The original data would be then queued to be sent down the socket 7.5 minutes from its original timestamp – that is, at least ~2.5 minutes from when its batch was fetched and at most 7.5 minutes since then. My question is this: is this the best way to approach the problem? Does this problem have any standard approaches or associated literature related to implementation best-practices and edge cases? I am a bit worried that the frequency of my fetches and the frequency in which the source is updated will get out of sync, leading to points where no data will be retrieved from the source. However, since my socket delay is greater than my fetch frequency, the subsequent fetch should retrieve newer data before the socket queue is empty. Is that correct? Am I missing something? Thanks!

    Read the article

  • ASP.NET MVC 3 (C#) Software Architecture

    - by ryanzec
    I am starting on a relatively large and ambitious ASP.NET MVC 3 project and just thinking about the best way to organize my code. The project is basically going to be a general management system that will be capable of supporting any type management system whether it be a blogging system, cms, reservation system, wikis, forums, project management system, etc…, each of them being just a separate 'module'. You can read more about it on my blog posted here : http://www.ryanzec.com/index.php/blog/details/8 (forgive me, the style of the site kinda sucks). For those who don't want to read the long blog post the basic idea is that the core system itself is nothing more than a users system with an admin interface to manage the users system. Then you just add on module as you need them and the module I will be creating is a simple blog post to test it out before I move on to the big module which is a project management system. Now I am just trying to think of the best way to structure this so that it is easy for users to add in there own modules but easy for me to update to core system without worrying about the user modifying the core code. I think the ideal way would be to have a number of core projects that user is specifically told not to modify otherwise the system may become unstable and future updates would not work. When the user wants to add in there own modules, they would just add in a new project (or multiple projects). The thing is I am not sure that it is even possible to use multiple projects all with their own controllers, razor view template, css, javascript, etc... in one web application. Ideally each module would have some of it own razor view templates, css, javascript, image files and also need access to some of the core razor view templates, css, javascript, image files which would is in a separate project. It is possible to have 1 web application run off of controllers, razor view templates, css, javascript, image files that are store in multiple projects? Is there a better was to structure this to allow the user to easily add in module with having to modify the core code?

    Read the article

  • How to approach scrum task burn down when tasks have multiple peoples involvement?

    - by AgileMan
    In my company, a single task can never be completed by one individual. There is going to be a separate person to QA and Code Review each task. What this means is that each individual will give their estimates, per task, as to how much time it will take to complete. The problem is, how should I approach burn down? If I aggregate the hours together, assume the following estimate: 10 hrs - Dev time 4 hrs - QA 4 hrs - Code Review. Task Estimate = 18hrs At the end of each day I ask that the task be updated with "how much time is left until it is done". However, each person generally just thinks about their part of it. Should they mark the effort remaining, and then ADD the effort estimates to that? How are you guys doing this? UPDATE To help clarify a few things, at my organization each Task within a story requires 3 people. Someone to develop the task. (do unit tests, ect...) A QA specialist to review task (they primarily do integration and regression tests) A Tech lead to do code review. I don't think there is a wrong way or a right way, but this is our way ... and that won't be changing. We work as a team to complete even the smallest level of a story whenever possible. You cannot actually test if something works until it is dev complete, and you cannot review the quality of the code either ... so the best you can do is split things up into small logical slices so that the bare minimum functionality can be tested and reviewed as early into the process as possible. My question to those that work this way would be how to burn down a "task" when they are setup this way. Unless a Task has it's own sub-tasks (which JIRA doesn't allow) ... I'm not sure the best way to accomplish tracking "what's left" on a daily basis.

    Read the article

  • DevWeek & SQL Social @ London

    - by Davide Mauri
    Yesterday I had my “SQL Server best practices for developers” session at DevWeek and I really enjoyed it a lot. For all those who asked, I’ll put slides and demos online as soon as possible. I’ve just waiting to know where I can put it (on my website or somewhere else), so it should be just a matter of some days. If you attended my session and would like to rate it, please use SpeakerRate here: http://speakerrate.com/talks/2857-sql-server-best-practices-for-developers I also have to thank Simon Sabin for the very nice event he organized for SQLSocial http://sqlblogcasts.com/blogs/simons/archive/2010/02/16/SQLSocial-presents-Itzik-Ben-gan--Greg-Low-and-Davide-Mauri.aspx A lot of people attended and we really had interesting discussions. And it was my first time doing a session at a pub, and I must say it's *really* funny and enjoyable, expecially when you have free beer :-) Now back to Italy to the “usual” work! Share this post: email it! | bookmark it! | digg it! | reddit! | kick it! | live it!

    Read the article

< Previous Page | 338 339 340 341 342 343 344 345 346 347 348 349  | Next Page >